Compare commits
13
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
341d276f18 | ||
|
|
2c8c98a11f | ||
|
|
bed7dd7f03 | ||
|
|
020d1ecf73 | ||
|
|
7d924b5c18 | ||
|
|
ad20eeb5c6 | ||
|
|
0ee456da94 | ||
|
|
d957a216a6 | ||
|
|
b4b4416c93 | ||
|
|
17deb91ca7 | ||
|
|
71e736b05f | ||
|
|
5703ed7ae0 | ||
|
|
57218959b0 |
@@ -0,0 +1,315 @@
|
||||
#!/usr/bin/env python3
|
||||
"""HTTP API for AIOS external-agent online task distribution.
|
||||
|
||||
The service is intended to bind to loopback and be published through the
|
||||
existing TLS reverse proxy. Control operations use a server-managed control
|
||||
Bearer token. Agents receive individual Bearer tokens at registration; only
|
||||
token hashes are persisted.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
import hmac
|
||||
import json
|
||||
import os
|
||||
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
|
||||
from pathlib import Path
|
||||
import re
|
||||
from typing import Any, Mapping
|
||||
from urllib.parse import urlparse
|
||||
|
||||
from aios_agent_task_protocol import TaskProtocolError, TaskStore
|
||||
|
||||
|
||||
MAX_BODY = 1024 * 1024
|
||||
TASK_ID_RE = re.compile(r"^[A-Za-z0-9][A-Za-z0-9._:-]{2,179}$")
|
||||
|
||||
|
||||
@dataclass(frozen=True, slots=True)
|
||||
class APIConfig:
|
||||
db_path: Path
|
||||
control_token_path: Path
|
||||
host: str = "127.0.0.1"
|
||||
port: int = 18940
|
||||
token_pepper: str = ""
|
||||
|
||||
@classmethod
|
||||
def from_env(cls) -> "APIConfig":
|
||||
return cls(
|
||||
db_path=Path(os.getenv(
|
||||
"AIOS_AGENT_TASK_DB",
|
||||
"/var/lib/aios-agent-task/state.sqlite3",
|
||||
)),
|
||||
control_token_path=Path(os.getenv(
|
||||
"AIOS_AGENT_TASK_CONTROL_TOKEN_FILE",
|
||||
"/run/credentials/aios-agent-task-api.service/control_token",
|
||||
)),
|
||||
host=os.getenv("AIOS_AGENT_TASK_HOST", "127.0.0.1"),
|
||||
port=int(os.getenv("AIOS_AGENT_TASK_PORT", "18940")),
|
||||
token_pepper=os.getenv("AIOS_AGENT_TASK_TOKEN_PEPPER", ""),
|
||||
)
|
||||
|
||||
|
||||
class AgentTaskApplication:
|
||||
def __init__(self, config: APIConfig) -> None:
|
||||
self.config = config
|
||||
self.store = TaskStore(
|
||||
config.db_path,
|
||||
token_pepper=config.token_pepper,
|
||||
)
|
||||
|
||||
def _control_token(self) -> str:
|
||||
path = self.config.control_token_path
|
||||
if path.is_symlink() or not path.is_file():
|
||||
raise TaskProtocolError("CONTROL_CREDENTIAL_UNAVAILABLE", 503)
|
||||
value = path.read_text(encoding="utf-8").strip()
|
||||
if len(value) < 32:
|
||||
raise TaskProtocolError("CONTROL_CREDENTIAL_INVALID", 503)
|
||||
return value
|
||||
|
||||
@staticmethod
|
||||
def _bearer(headers: Mapping[str, str]) -> str:
|
||||
value = str(headers.get("Authorization") or "")
|
||||
if not value.startswith("Bearer "):
|
||||
return ""
|
||||
return value[7:].strip()
|
||||
|
||||
def authenticate_control(self, headers: Mapping[str, str]) -> None:
|
||||
supplied = self._bearer(headers)
|
||||
expected = self._control_token()
|
||||
if not supplied or not hmac.compare_digest(supplied, expected):
|
||||
raise TaskProtocolError("UNAUTHORIZED", 401)
|
||||
|
||||
def authenticate_agent(self, headers: Mapping[str, str]) -> dict[str, Any]:
|
||||
return self.store.authenticate_agent(self._bearer(headers))
|
||||
|
||||
def principal(self, headers: Mapping[str, str]) -> tuple[str, dict[str, Any] | None]:
|
||||
supplied = self._bearer(headers)
|
||||
if not supplied:
|
||||
raise TaskProtocolError("UNAUTHORIZED", 401)
|
||||
try:
|
||||
expected = self._control_token()
|
||||
except TaskProtocolError:
|
||||
expected = ""
|
||||
if expected and hmac.compare_digest(supplied, expected):
|
||||
return "control", None
|
||||
return "agent", self.store.authenticate_agent(supplied)
|
||||
|
||||
def health(self) -> dict[str, Any]:
|
||||
return {
|
||||
"status": "ok",
|
||||
"schema": "aios.agent-task.v1",
|
||||
"task_states": [
|
||||
"QUEUED", "CLAIMED", "RUNNING", "VERIFYING",
|
||||
"SUCCEEDED", "FAILED", "EXPIRED", "CANCELLED",
|
||||
],
|
||||
"lease_recovery": True,
|
||||
"atomic_claim": True,
|
||||
"idempotent_result_submission": True,
|
||||
"server_side_verification_required": True,
|
||||
"plaintext_credentials_persisted": False,
|
||||
}
|
||||
|
||||
def register_agent(self, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
self.authenticate_control(headers)
|
||||
return self.store.register_agent(
|
||||
agent_id=str(body.get("agent_id") or ""),
|
||||
agent_type=str(body.get("agent_type") or ""),
|
||||
agent_version=str(body.get("agent_version") or ""),
|
||||
capabilities=body.get("capabilities") or [],
|
||||
metadata=body.get("metadata") if isinstance(body.get("metadata"), Mapping) else {},
|
||||
)
|
||||
|
||||
def create_task(self, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
self.authenticate_control(headers)
|
||||
return self.store.create_task(
|
||||
task_id=str(body.get("task_id") or "") or None,
|
||||
goal=str(body.get("goal") or ""),
|
||||
instructions=str(body.get("instructions") or ""),
|
||||
success_criteria=body.get("success_criteria"),
|
||||
required_capabilities=body.get("required_capabilities") or [],
|
||||
constraints=body.get("constraints") if isinstance(body.get("constraints"), Mapping) else {},
|
||||
metadata=body.get("metadata") if isinstance(body.get("metadata"), Mapping) else {},
|
||||
deadline=str(body.get("deadline") or "") or None,
|
||||
priority=body.get("priority", 0),
|
||||
max_attempts=body.get("max_attempts", 3),
|
||||
idempotency_key=str(body.get("idempotency_key") or ""),
|
||||
)
|
||||
|
||||
def poll(self, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
agent = self.authenticate_agent(headers)
|
||||
tasks = self.store.poll_tasks(
|
||||
agent_id=agent["agent_id"],
|
||||
limit=int(body.get("limit", 20)),
|
||||
)
|
||||
return {"agent_id": agent["agent_id"], "tasks": tasks}
|
||||
|
||||
def claim(self, task_id: str, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
agent = self.authenticate_agent(headers)
|
||||
return self.store.claim_task(
|
||||
task_id=task_id,
|
||||
agent_id=agent["agent_id"],
|
||||
lease_seconds=int(body.get("lease_seconds", 300)),
|
||||
)
|
||||
|
||||
def heartbeat(self, task_id: str, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
agent = self.authenticate_agent(headers)
|
||||
return self.store.heartbeat(
|
||||
task_id=task_id,
|
||||
agent_id=agent["agent_id"],
|
||||
lease_token=str(body.get("lease_token") or ""),
|
||||
progress=body.get("progress") if isinstance(body.get("progress"), Mapping) else {},
|
||||
lease_seconds=int(body.get("lease_seconds", 300)),
|
||||
)
|
||||
|
||||
def result(self, task_id: str, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
agent = self.authenticate_agent(headers)
|
||||
result = body.get("result")
|
||||
if not isinstance(result, Mapping):
|
||||
raise TaskProtocolError("RESULT_INVALID", 400)
|
||||
return self.store.submit_result(
|
||||
task_id=task_id,
|
||||
agent_id=agent["agent_id"],
|
||||
lease_token=str(body.get("lease_token") or ""),
|
||||
idempotency_key=str(body.get("idempotency_key") or ""),
|
||||
result=result,
|
||||
)
|
||||
|
||||
def verify(self, task_id: str, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
self.authenticate_control(headers)
|
||||
return self.store.verify_task(
|
||||
task_id=task_id,
|
||||
decision=str(body.get("decision") or ""),
|
||||
verification=body.get("verification") if isinstance(body.get("verification"), Mapping) else {},
|
||||
)
|
||||
|
||||
def cancel(self, task_id: str, headers: Mapping[str, str], body: Mapping[str, Any]) -> dict[str, Any]:
|
||||
self.authenticate_control(headers)
|
||||
return self.store.cancel_task(task_id=task_id, reason=str(body.get("reason") or ""))
|
||||
|
||||
def get_task(self, task_id: str, headers: Mapping[str, str]) -> dict[str, Any]:
|
||||
kind, agent = self.principal(headers)
|
||||
task = self.store.get_task(task_id)
|
||||
if kind == "agent" and task.get("claimed_by") != agent["agent_id"]:
|
||||
raise TaskProtocolError("FORBIDDEN", 403)
|
||||
return task
|
||||
|
||||
def events(self, task_id: str, headers: Mapping[str, str]) -> dict[str, Any]:
|
||||
self.authenticate_control(headers)
|
||||
return {"task_id": task_id, "events": self.store.get_events(task_id)}
|
||||
|
||||
|
||||
class Handler(BaseHTTPRequestHandler):
|
||||
app: AgentTaskApplication | None = None
|
||||
server_version = "AIOSAgentTaskAPI/1"
|
||||
|
||||
def log_message(self, _format: str, *_args: Any) -> None:
|
||||
return
|
||||
|
||||
def _headers(self) -> dict[str, str]:
|
||||
return {key: value for key, value in self.headers.items()}
|
||||
|
||||
def _json(self, value: Any, status: int = 200) -> None:
|
||||
payload = json.dumps(value, ensure_ascii=False, sort_keys=True).encode("utf-8")
|
||||
self.send_response(status)
|
||||
self.send_header("Content-Type", "application/json; charset=utf-8")
|
||||
self.send_header("Cache-Control", "no-store")
|
||||
self.send_header("Content-Length", str(len(payload)))
|
||||
self.end_headers()
|
||||
self.wfile.write(payload)
|
||||
|
||||
def _body(self) -> dict[str, Any]:
|
||||
try:
|
||||
size = int(self.headers.get("Content-Length", "0"))
|
||||
except ValueError as exc:
|
||||
raise TaskProtocolError("CONTENT_LENGTH_INVALID", 400) from exc
|
||||
if size < 0 or size > MAX_BODY:
|
||||
raise TaskProtocolError("REQUEST_TOO_LARGE", 413)
|
||||
raw = self.rfile.read(size) if size else b"{}"
|
||||
try:
|
||||
value = json.loads(raw.decode("utf-8"))
|
||||
except (UnicodeDecodeError, json.JSONDecodeError) as exc:
|
||||
raise TaskProtocolError("JSON_INVALID", 400) from exc
|
||||
if not isinstance(value, dict):
|
||||
raise TaskProtocolError("JSON_OBJECT_REQUIRED", 400)
|
||||
return value
|
||||
|
||||
@staticmethod
|
||||
def _task_route(path: str) -> tuple[str, str | None]:
|
||||
match = re.fullmatch(
|
||||
r"/v1/tasks/([A-Za-z0-9][A-Za-z0-9._:-]{2,179})(?:/(claim|heartbeat|result|verify|cancel|events))?",
|
||||
path,
|
||||
)
|
||||
if match:
|
||||
return match.group(1), match.group(2)
|
||||
return "", None
|
||||
|
||||
def _run(self, callback) -> None:
|
||||
try:
|
||||
callback()
|
||||
except TaskProtocolError as exc:
|
||||
self._json({"error": exc.code}, exc.http_status)
|
||||
except Exception as exc:
|
||||
self._json({"error": "INTERNAL_ERROR", "detail": type(exc).__name__}, 500)
|
||||
|
||||
def do_GET(self) -> None:
|
||||
def action() -> None:
|
||||
if self.app is None:
|
||||
raise TaskProtocolError("SERVICE_UNAVAILABLE", 503)
|
||||
path = urlparse(self.path).path.rstrip("/") or "/"
|
||||
if path == "/health":
|
||||
self._json(self.app.health())
|
||||
return
|
||||
task_id, suffix = self._task_route(path)
|
||||
if task_id and suffix is None:
|
||||
self._json(self.app.get_task(task_id, self._headers()))
|
||||
return
|
||||
if task_id and suffix == "events":
|
||||
self._json(self.app.events(task_id, self._headers()))
|
||||
return
|
||||
self._json({"error": "NOT_FOUND"}, 404)
|
||||
self._run(action)
|
||||
|
||||
def do_POST(self) -> None:
|
||||
def action() -> None:
|
||||
if self.app is None:
|
||||
raise TaskProtocolError("SERVICE_UNAVAILABLE", 503)
|
||||
path = urlparse(self.path).path.rstrip("/") or "/"
|
||||
body = self._body()
|
||||
if path == "/v1/agents/register":
|
||||
self._json(self.app.register_agent(self._headers(), body), 201)
|
||||
return
|
||||
if path == "/v1/tasks":
|
||||
self._json(self.app.create_task(self._headers(), body), 201)
|
||||
return
|
||||
if path == "/v1/tasks/poll":
|
||||
self._json(self.app.poll(self._headers(), body))
|
||||
return
|
||||
task_id, suffix = self._task_route(path)
|
||||
if not task_id:
|
||||
self._json({"error": "NOT_FOUND"}, 404)
|
||||
return
|
||||
if suffix == "claim":
|
||||
self._json(self.app.claim(task_id, self._headers(), body))
|
||||
elif suffix == "heartbeat":
|
||||
self._json(self.app.heartbeat(task_id, self._headers(), body))
|
||||
elif suffix == "result":
|
||||
self._json(self.app.result(task_id, self._headers(), body))
|
||||
elif suffix == "verify":
|
||||
self._json(self.app.verify(task_id, self._headers(), body))
|
||||
elif suffix == "cancel":
|
||||
self._json(self.app.cancel(task_id, self._headers(), body))
|
||||
else:
|
||||
self._json({"error": "NOT_FOUND"}, 404)
|
||||
self._run(action)
|
||||
|
||||
|
||||
def serve(config: APIConfig | None = None) -> None:
|
||||
config = config or APIConfig.from_env()
|
||||
application = AgentTaskApplication(config)
|
||||
Handler.app = application
|
||||
ThreadingHTTPServer((config.host, config.port), Handler).serve_forever()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
serve()
|
||||
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,870 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Durable online task protocol for external AIOS agents.
|
||||
|
||||
The store is deliberately independent from Git/Gitea mutation. It provides a
|
||||
small durable state machine for publishing work, atomic claim/lease ownership,
|
||||
heartbeat, idempotent result submission, explicit server-side verification,
|
||||
lease recovery, and audit events.
|
||||
|
||||
No plaintext agent credential or task lease token is persisted.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime, timedelta, timezone
|
||||
import hashlib
|
||||
import hmac
|
||||
import json
|
||||
from pathlib import Path
|
||||
import secrets
|
||||
import sqlite3
|
||||
from typing import Any, Callable, Iterable, Mapping
|
||||
import uuid
|
||||
|
||||
from aios_capability_contract import has_capabilities
|
||||
from aios_online_agent_execution import (
|
||||
ExecutionProjectionError,
|
||||
validate_execution_envelope,
|
||||
)
|
||||
|
||||
|
||||
TERMINAL_STATES = frozenset({"SUCCEEDED", "FAILED", "EXPIRED", "CANCELLED"})
|
||||
ACTIVE_LEASE_STATES = frozenset({"CLAIMED", "RUNNING"})
|
||||
TASK_STATES = frozenset({
|
||||
"QUEUED", "CLAIMED", "RUNNING", "VERIFYING",
|
||||
"SUCCEEDED", "FAILED", "EXPIRED", "CANCELLED",
|
||||
})
|
||||
VERIFY_DECISIONS = frozenset({"SUCCEEDED", "NEEDS_REWORK", "FAILED"})
|
||||
|
||||
|
||||
GOVERNED_WRITE_CAPABILITIES = frozenset({
|
||||
"governed.write.branch",
|
||||
"governed.write.file",
|
||||
})
|
||||
EXECUTION_PROJECTION_SCHEMA = "aios.online-agent.task-projection.v1"
|
||||
|
||||
|
||||
def _validate_write_execution_binding(
|
||||
task_id: str,
|
||||
required_capabilities: list[str],
|
||||
metadata: Mapping[str, Any],
|
||||
) -> None:
|
||||
required = set(required_capabilities)
|
||||
if not required.intersection(GOVERNED_WRITE_CAPABILITIES):
|
||||
return
|
||||
execution = metadata.get("execution")
|
||||
projection = metadata.get("execution_projection")
|
||||
if not isinstance(execution, Mapping) or not isinstance(projection, Mapping):
|
||||
raise TaskProtocolError("GOVERNED_EXECUTION_BINDING_REQUIRED", 409)
|
||||
try:
|
||||
canonical = validate_execution_envelope(dict(execution))
|
||||
except (ExecutionProjectionError, TypeError, ValueError, KeyError) as exc:
|
||||
raise TaskProtocolError("EXECUTION_ENVELOPE_INVALID", 409) from exc
|
||||
if canonical.get("task_id") != task_id:
|
||||
raise TaskProtocolError("TASK_BINDING_CONFLICT", 409)
|
||||
if set(canonical.get("capability_scope") or ()) != required:
|
||||
raise TaskProtocolError("EXECUTION_CAPABILITY_SCOPE_CONFLICT", 409)
|
||||
if (
|
||||
projection.get("schema") != EXECUTION_PROJECTION_SCHEMA
|
||||
or projection.get("task_id") != task_id
|
||||
or projection.get("job_id") != canonical.get("job_id")
|
||||
or projection.get("execution_digest") != canonical.get("execution_digest")
|
||||
):
|
||||
raise TaskProtocolError("EXECUTION_PROJECTION_INVALID", 409)
|
||||
|
||||
|
||||
class TaskProtocolError(RuntimeError):
|
||||
def __init__(self, code: str, http_status: int = 409) -> None:
|
||||
super().__init__(code)
|
||||
self.code = code
|
||||
self.http_status = http_status
|
||||
|
||||
|
||||
@dataclass(frozen=True, slots=True)
|
||||
class Principal:
|
||||
kind: str
|
||||
identifier: str
|
||||
|
||||
|
||||
def utcnow() -> datetime:
|
||||
return datetime.now(timezone.utc)
|
||||
|
||||
|
||||
def iso(value: datetime) -> str:
|
||||
return value.astimezone(timezone.utc).isoformat()
|
||||
|
||||
|
||||
def parse_time(value: str | None) -> datetime | None:
|
||||
if not value:
|
||||
return None
|
||||
parsed = datetime.fromisoformat(value)
|
||||
if parsed.tzinfo is None:
|
||||
parsed = parsed.replace(tzinfo=timezone.utc)
|
||||
return parsed.astimezone(timezone.utc)
|
||||
|
||||
|
||||
def canonical_json(value: Any) -> str:
|
||||
return json.dumps(value, ensure_ascii=False, sort_keys=True, separators=(",", ":"))
|
||||
|
||||
|
||||
def digest_json(value: Any) -> str:
|
||||
return hashlib.sha256(canonical_json(value).encode("utf-8")).hexdigest()
|
||||
|
||||
|
||||
def _string_list(value: Iterable[object] | None, field: str, *, max_items: int = 128) -> list[str]:
|
||||
if value is None:
|
||||
return []
|
||||
if isinstance(value, (str, bytes)) or not isinstance(value, Iterable):
|
||||
raise TaskProtocolError(field.upper() + "_INVALID", 400)
|
||||
result: list[str] = []
|
||||
for item in value:
|
||||
text = str(item).strip()
|
||||
if not text or len(text) > 160:
|
||||
raise TaskProtocolError(field.upper() + "_INVALID", 400)
|
||||
result.append(text)
|
||||
if len(result) > max_items:
|
||||
raise TaskProtocolError(field.upper() + "_TOO_LARGE", 400)
|
||||
return sorted(set(result))
|
||||
|
||||
|
||||
def _json_object(value: object | None, field: str, *, max_bytes: int = 128 * 1024) -> dict[str, Any]:
|
||||
if value is None:
|
||||
return {}
|
||||
if not isinstance(value, Mapping):
|
||||
raise TaskProtocolError(field.upper() + "_INVALID", 400)
|
||||
result = dict(value)
|
||||
if len(canonical_json(result).encode("utf-8")) > max_bytes:
|
||||
raise TaskProtocolError(field.upper() + "_TOO_LARGE", 413)
|
||||
return result
|
||||
|
||||
|
||||
def _criteria(value: object | None) -> list[Any]:
|
||||
if value is None:
|
||||
return []
|
||||
if isinstance(value, (str, bytes)) or not isinstance(value, list):
|
||||
raise TaskProtocolError("SUCCESS_CRITERIA_INVALID", 400)
|
||||
if len(value) > 128:
|
||||
raise TaskProtocolError("SUCCESS_CRITERIA_TOO_LARGE", 413)
|
||||
if len(canonical_json(value).encode("utf-8")) > 128 * 1024:
|
||||
raise TaskProtocolError("SUCCESS_CRITERIA_TOO_LARGE", 413)
|
||||
return list(value)
|
||||
|
||||
|
||||
class TaskStore:
|
||||
"""SQLite-backed durable task store.
|
||||
|
||||
SQLite BEGIN IMMEDIATE serializes claims so two agents cannot successfully
|
||||
claim the same queued task. WAL mode allows concurrent readers.
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
path: str | Path,
|
||||
*,
|
||||
now: Callable[[], datetime] = utcnow,
|
||||
token_pepper: str = "",
|
||||
) -> None:
|
||||
self.path = Path(path)
|
||||
self.path.parent.mkdir(parents=True, exist_ok=True)
|
||||
self.now = now
|
||||
self.token_pepper = token_pepper
|
||||
self._initialize()
|
||||
|
||||
def _connect(self) -> sqlite3.Connection:
|
||||
connection = sqlite3.connect(self.path, timeout=10, isolation_level=None)
|
||||
connection.row_factory = sqlite3.Row
|
||||
connection.execute("PRAGMA foreign_keys=ON")
|
||||
connection.execute("PRAGMA journal_mode=WAL")
|
||||
connection.execute("PRAGMA synchronous=FULL")
|
||||
connection.execute("PRAGMA busy_timeout=10000")
|
||||
return connection
|
||||
|
||||
def _initialize(self) -> None:
|
||||
with self._connect() as connection:
|
||||
connection.executescript(
|
||||
"""
|
||||
CREATE TABLE IF NOT EXISTS agents (
|
||||
agent_id TEXT PRIMARY KEY,
|
||||
agent_type TEXT NOT NULL,
|
||||
agent_version TEXT NOT NULL,
|
||||
capabilities_json TEXT NOT NULL,
|
||||
metadata_json TEXT NOT NULL,
|
||||
token_hash TEXT NOT NULL UNIQUE,
|
||||
status TEXT NOT NULL,
|
||||
created_at TEXT NOT NULL,
|
||||
last_seen_at TEXT NOT NULL
|
||||
);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS tasks (
|
||||
task_id TEXT PRIMARY KEY,
|
||||
goal TEXT NOT NULL,
|
||||
instructions TEXT NOT NULL,
|
||||
success_criteria_json TEXT NOT NULL,
|
||||
required_capabilities_json TEXT NOT NULL,
|
||||
constraints_json TEXT NOT NULL,
|
||||
metadata_json TEXT NOT NULL,
|
||||
deadline TEXT,
|
||||
priority INTEGER NOT NULL,
|
||||
max_attempts INTEGER NOT NULL,
|
||||
attempt INTEGER NOT NULL DEFAULT 0,
|
||||
status TEXT NOT NULL,
|
||||
claimed_by TEXT,
|
||||
lease_token_hash TEXT,
|
||||
lease_expires_at TEXT,
|
||||
result_json TEXT,
|
||||
result_digest TEXT,
|
||||
result_idempotency_key TEXT,
|
||||
verification_json TEXT,
|
||||
revision INTEGER NOT NULL DEFAULT 1,
|
||||
created_at TEXT NOT NULL,
|
||||
updated_at TEXT NOT NULL,
|
||||
FOREIGN KEY(claimed_by) REFERENCES agents(agent_id)
|
||||
);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS idempotency (
|
||||
scope TEXT NOT NULL,
|
||||
key TEXT NOT NULL,
|
||||
payload_digest TEXT NOT NULL,
|
||||
response_json TEXT NOT NULL,
|
||||
created_at TEXT NOT NULL,
|
||||
PRIMARY KEY(scope, key)
|
||||
);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS events (
|
||||
event_id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
task_id TEXT,
|
||||
agent_id TEXT,
|
||||
event_type TEXT NOT NULL,
|
||||
payload_json TEXT NOT NULL,
|
||||
created_at TEXT NOT NULL
|
||||
);
|
||||
|
||||
CREATE INDEX IF NOT EXISTS idx_tasks_queue
|
||||
ON tasks(status, priority DESC, created_at);
|
||||
|
||||
CREATE INDEX IF NOT EXISTS idx_tasks_lease
|
||||
ON tasks(status, lease_expires_at);
|
||||
|
||||
CREATE INDEX IF NOT EXISTS idx_events_task
|
||||
ON events(task_id, event_id);
|
||||
"""
|
||||
)
|
||||
|
||||
def _token_hash(self, token: str) -> str:
|
||||
return hashlib.sha256((self.token_pepper + "\0" + token).encode("utf-8")).hexdigest()
|
||||
|
||||
@staticmethod
|
||||
def _row_task(row: sqlite3.Row, *, include_result: bool = True) -> dict[str, Any]:
|
||||
value: dict[str, Any] = {
|
||||
"task_id": row["task_id"],
|
||||
"goal": row["goal"],
|
||||
"instructions": row["instructions"],
|
||||
"success_criteria": json.loads(row["success_criteria_json"]),
|
||||
"required_capabilities": json.loads(row["required_capabilities_json"]),
|
||||
"constraints": json.loads(row["constraints_json"]),
|
||||
"metadata": json.loads(row["metadata_json"]),
|
||||
"deadline": row["deadline"],
|
||||
"priority": row["priority"],
|
||||
"max_attempts": row["max_attempts"],
|
||||
"attempt": row["attempt"],
|
||||
"status": row["status"],
|
||||
"claimed_by": row["claimed_by"],
|
||||
"lease_expires_at": row["lease_expires_at"],
|
||||
"revision": row["revision"],
|
||||
"created_at": row["created_at"],
|
||||
"updated_at": row["updated_at"],
|
||||
}
|
||||
if include_result:
|
||||
value["result"] = json.loads(row["result_json"]) if row["result_json"] else None
|
||||
value["verification"] = (
|
||||
json.loads(row["verification_json"]) if row["verification_json"] else None
|
||||
)
|
||||
return value
|
||||
|
||||
def _event(
|
||||
self,
|
||||
connection: sqlite3.Connection,
|
||||
event_type: str,
|
||||
*,
|
||||
task_id: str | None = None,
|
||||
agent_id: str | None = None,
|
||||
payload: Mapping[str, Any] | None = None,
|
||||
) -> None:
|
||||
connection.execute(
|
||||
"INSERT INTO events(task_id,agent_id,event_type,payload_json,created_at)"
|
||||
" VALUES(?,?,?,?,?)",
|
||||
(task_id, agent_id, event_type, canonical_json(dict(payload or {})), iso(self.now())),
|
||||
)
|
||||
|
||||
def register_agent(
|
||||
self,
|
||||
*,
|
||||
agent_id: str,
|
||||
agent_type: str,
|
||||
agent_version: str,
|
||||
capabilities: Iterable[object],
|
||||
metadata: Mapping[str, Any] | None = None,
|
||||
) -> dict[str, Any]:
|
||||
agent_id = agent_id.strip()
|
||||
agent_type = agent_type.strip()
|
||||
agent_version = agent_version.strip()
|
||||
if not agent_id or len(agent_id) > 160 or not agent_type or not agent_version:
|
||||
raise TaskProtocolError("AGENT_IDENTITY_INVALID", 400)
|
||||
caps = _string_list(capabilities, "capabilities")
|
||||
meta = _json_object(metadata, "metadata")
|
||||
token = "agt_" + secrets.token_urlsafe(32)
|
||||
now = iso(self.now())
|
||||
try:
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
connection.execute(
|
||||
"INSERT INTO agents(agent_id,agent_type,agent_version,capabilities_json,"
|
||||
"metadata_json,token_hash,status,created_at,last_seen_at)"
|
||||
" VALUES(?,?,?,?,?,?,?,?,?)",
|
||||
(
|
||||
agent_id,
|
||||
agent_type,
|
||||
agent_version,
|
||||
canonical_json(caps),
|
||||
canonical_json(meta),
|
||||
self._token_hash(token),
|
||||
"ACTIVE",
|
||||
now,
|
||||
now,
|
||||
),
|
||||
)
|
||||
self._event(
|
||||
connection,
|
||||
"AGENT_REGISTERED",
|
||||
agent_id=agent_id,
|
||||
payload={"agent_type": agent_type, "agent_version": agent_version},
|
||||
)
|
||||
connection.commit()
|
||||
except sqlite3.IntegrityError as exc:
|
||||
raise TaskProtocolError("AGENT_ALREADY_EXISTS", 409) from exc
|
||||
return {
|
||||
"agent_id": agent_id,
|
||||
"agent_type": agent_type,
|
||||
"agent_version": agent_version,
|
||||
"capabilities": caps,
|
||||
"agent_token": token,
|
||||
"credential_returned_once": True,
|
||||
}
|
||||
|
||||
def authenticate_agent(self, token: str) -> dict[str, Any]:
|
||||
if not token:
|
||||
raise TaskProtocolError("UNAUTHORIZED", 401)
|
||||
token_hash = self._token_hash(token)
|
||||
with self._connect() as connection:
|
||||
row = connection.execute(
|
||||
"SELECT * FROM agents WHERE token_hash=? AND status='ACTIVE'",
|
||||
(token_hash,),
|
||||
).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("UNAUTHORIZED", 401)
|
||||
connection.execute(
|
||||
"UPDATE agents SET last_seen_at=? WHERE agent_id=?",
|
||||
(iso(self.now()), row["agent_id"]),
|
||||
)
|
||||
return {
|
||||
"agent_id": row["agent_id"],
|
||||
"agent_type": row["agent_type"],
|
||||
"agent_version": row["agent_version"],
|
||||
"capabilities": json.loads(row["capabilities_json"]),
|
||||
"metadata": json.loads(row["metadata_json"]),
|
||||
}
|
||||
|
||||
def _idempotent_lookup(
|
||||
self,
|
||||
connection: sqlite3.Connection,
|
||||
*,
|
||||
scope: str,
|
||||
key: str,
|
||||
payload_digest: str,
|
||||
) -> dict[str, Any] | None:
|
||||
row = connection.execute(
|
||||
"SELECT payload_digest,response_json FROM idempotency WHERE scope=? AND key=?",
|
||||
(scope, key),
|
||||
).fetchone()
|
||||
if row is None:
|
||||
return None
|
||||
if not hmac.compare_digest(row["payload_digest"], payload_digest):
|
||||
raise TaskProtocolError("IDEMPOTENCY_CONFLICT", 409)
|
||||
return json.loads(row["response_json"])
|
||||
|
||||
def _idempotent_store(
|
||||
self,
|
||||
connection: sqlite3.Connection,
|
||||
*,
|
||||
scope: str,
|
||||
key: str,
|
||||
payload_digest: str,
|
||||
response: Mapping[str, Any],
|
||||
) -> None:
|
||||
connection.execute(
|
||||
"INSERT INTO idempotency(scope,key,payload_digest,response_json,created_at)"
|
||||
" VALUES(?,?,?,?,?)",
|
||||
(scope, key, payload_digest, canonical_json(dict(response)), iso(self.now())),
|
||||
)
|
||||
|
||||
def create_task(
|
||||
self,
|
||||
*,
|
||||
task_id: str | None,
|
||||
goal: str,
|
||||
instructions: str,
|
||||
success_criteria: object | None,
|
||||
required_capabilities: Iterable[object] | None = None,
|
||||
constraints: Mapping[str, Any] | None = None,
|
||||
metadata: Mapping[str, Any] | None = None,
|
||||
deadline: str | None = None,
|
||||
priority: int = 0,
|
||||
max_attempts: int = 3,
|
||||
idempotency_key: str,
|
||||
) -> dict[str, Any]:
|
||||
task_id = (task_id or ("task-" + uuid.uuid4().hex)).strip()
|
||||
goal = goal.strip()
|
||||
instructions = instructions.strip()
|
||||
idempotency_key = idempotency_key.strip()
|
||||
if not task_id or len(task_id) > 180 or not goal or not instructions:
|
||||
raise TaskProtocolError("TASK_BINDING_INVALID", 400)
|
||||
if not idempotency_key or len(idempotency_key) > 160:
|
||||
raise TaskProtocolError("IDEMPOTENCY_KEY_INVALID", 400)
|
||||
if not isinstance(priority, int) or isinstance(priority, bool) or not -100 <= priority <= 100:
|
||||
raise TaskProtocolError("PRIORITY_INVALID", 400)
|
||||
if not isinstance(max_attempts, int) or isinstance(max_attempts, bool) or not 1 <= max_attempts <= 20:
|
||||
raise TaskProtocolError("MAX_ATTEMPTS_INVALID", 400)
|
||||
parsed_deadline = parse_time(deadline)
|
||||
if parsed_deadline and parsed_deadline <= self.now():
|
||||
raise TaskProtocolError("DEADLINE_INVALID", 400)
|
||||
criteria = _criteria(success_criteria)
|
||||
required = _string_list(required_capabilities, "required_capabilities")
|
||||
constraints_value = _json_object(constraints, "constraints")
|
||||
metadata_value = _json_object(metadata, "metadata")
|
||||
_validate_write_execution_binding(task_id, required, metadata_value)
|
||||
payload = {
|
||||
"task_id": task_id,
|
||||
"goal": goal,
|
||||
"instructions": instructions,
|
||||
"success_criteria": criteria,
|
||||
"required_capabilities": required,
|
||||
"constraints": constraints_value,
|
||||
"metadata": metadata_value,
|
||||
"deadline": iso(parsed_deadline) if parsed_deadline else None,
|
||||
"priority": priority,
|
||||
"max_attempts": max_attempts,
|
||||
}
|
||||
payload_digest = digest_json(payload)
|
||||
now = iso(self.now())
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
replay = self._idempotent_lookup(
|
||||
connection,
|
||||
scope="create_task",
|
||||
key=idempotency_key,
|
||||
payload_digest=payload_digest,
|
||||
)
|
||||
if replay is not None:
|
||||
connection.commit()
|
||||
return {**replay, "idempotent": True}
|
||||
try:
|
||||
connection.execute(
|
||||
"INSERT INTO tasks(task_id,goal,instructions,success_criteria_json,"
|
||||
"required_capabilities_json,constraints_json,metadata_json,deadline,priority,"
|
||||
"max_attempts,attempt,status,revision,created_at,updated_at)"
|
||||
" VALUES(?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)",
|
||||
(
|
||||
task_id,
|
||||
goal,
|
||||
instructions,
|
||||
canonical_json(criteria),
|
||||
canonical_json(required),
|
||||
canonical_json(constraints_value),
|
||||
canonical_json(metadata_value),
|
||||
payload["deadline"],
|
||||
priority,
|
||||
max_attempts,
|
||||
0,
|
||||
"QUEUED",
|
||||
1,
|
||||
now,
|
||||
now,
|
||||
),
|
||||
)
|
||||
except sqlite3.IntegrityError as exc:
|
||||
raise TaskProtocolError("TASK_ID_CONFLICT", 409) from exc
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
result = self._row_task(row)
|
||||
self._idempotent_store(
|
||||
connection,
|
||||
scope="create_task",
|
||||
key=idempotency_key,
|
||||
payload_digest=payload_digest,
|
||||
response=result,
|
||||
)
|
||||
self._event(connection, "TASK_CREATED", task_id=task_id, payload={"priority": priority})
|
||||
connection.commit()
|
||||
return result
|
||||
|
||||
def _expire_locked(self, connection: sqlite3.Connection) -> None:
|
||||
now_dt = self.now()
|
||||
now = iso(now_dt)
|
||||
rows = connection.execute(
|
||||
"SELECT * FROM tasks WHERE status IN ('QUEUED','CLAIMED','RUNNING')"
|
||||
).fetchall()
|
||||
for row in rows:
|
||||
deadline = parse_time(row["deadline"])
|
||||
if deadline is not None and deadline <= now_dt:
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status='EXPIRED',lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,revision=revision+1,updated_at=? WHERE task_id=?",
|
||||
(now, row["task_id"]),
|
||||
)
|
||||
self._event(connection, "TASK_EXPIRED", task_id=row["task_id"])
|
||||
continue
|
||||
if row["status"] in ACTIVE_LEASE_STATES:
|
||||
lease_expires = parse_time(row["lease_expires_at"])
|
||||
if lease_expires is not None and lease_expires <= now_dt:
|
||||
if row["attempt"] >= row["max_attempts"]:
|
||||
next_state = "FAILED"
|
||||
else:
|
||||
next_state = "QUEUED"
|
||||
if next_state == "FAILED":
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status=?,lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,revision=revision+1,updated_at=? WHERE task_id=?",
|
||||
(next_state, now, row["task_id"]),
|
||||
)
|
||||
else:
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status=?,claimed_by=NULL,lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,revision=revision+1,updated_at=? WHERE task_id=?",
|
||||
(next_state, now, row["task_id"]),
|
||||
)
|
||||
self._event(
|
||||
connection,
|
||||
"LEASE_EXPIRED",
|
||||
task_id=row["task_id"],
|
||||
agent_id=row["claimed_by"],
|
||||
payload={"next_state": next_state, "attempt": row["attempt"]},
|
||||
)
|
||||
|
||||
def expire_leases(self) -> int:
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
before = connection.total_changes
|
||||
self._expire_locked(connection)
|
||||
changed = connection.total_changes - before
|
||||
connection.commit()
|
||||
return changed
|
||||
|
||||
def poll_tasks(self, *, agent_id: str, limit: int = 20) -> list[dict[str, Any]]:
|
||||
if not 1 <= int(limit) <= 100:
|
||||
raise TaskProtocolError("LIMIT_INVALID", 400)
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
self._expire_locked(connection)
|
||||
agent = connection.execute(
|
||||
"SELECT capabilities_json FROM agents WHERE agent_id=? AND status='ACTIVE'",
|
||||
(agent_id,),
|
||||
).fetchone()
|
||||
if agent is None:
|
||||
raise TaskProtocolError("AGENT_NOT_FOUND", 404)
|
||||
capabilities = json.loads(agent["capabilities_json"])
|
||||
rows = connection.execute(
|
||||
"SELECT * FROM tasks WHERE status='QUEUED' ORDER BY priority DESC,created_at ASC"
|
||||
).fetchall()
|
||||
result = []
|
||||
for row in rows:
|
||||
required = json.loads(row["required_capabilities_json"])
|
||||
if has_capabilities(capabilities, required):
|
||||
result.append(self._row_task(row, include_result=False))
|
||||
if len(result) >= int(limit):
|
||||
break
|
||||
connection.commit()
|
||||
return result
|
||||
|
||||
def _agent_capabilities(self, connection: sqlite3.Connection, agent_id: str) -> set[str]:
|
||||
row = connection.execute(
|
||||
"SELECT capabilities_json FROM agents WHERE agent_id=? AND status='ACTIVE'",
|
||||
(agent_id,),
|
||||
).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("AGENT_NOT_FOUND", 404)
|
||||
return set(json.loads(row["capabilities_json"]))
|
||||
|
||||
def claim_task(
|
||||
self,
|
||||
*,
|
||||
task_id: str,
|
||||
agent_id: str,
|
||||
lease_seconds: int = 300,
|
||||
) -> dict[str, Any]:
|
||||
if not isinstance(lease_seconds, int) or isinstance(lease_seconds, bool) or not 30 <= lease_seconds <= 3600:
|
||||
raise TaskProtocolError("LEASE_SECONDS_INVALID", 400)
|
||||
lease_token = "lease_" + secrets.token_urlsafe(32)
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
self._expire_locked(connection)
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
if row["status"] != "QUEUED":
|
||||
raise TaskProtocolError("TASK_NOT_CLAIMABLE", 409)
|
||||
capabilities = self._agent_capabilities(connection, agent_id)
|
||||
required = json.loads(row["required_capabilities_json"])
|
||||
if not has_capabilities(capabilities, required):
|
||||
raise TaskProtocolError("AGENT_CAPABILITY_MISMATCH", 403)
|
||||
attempt = int(row["attempt"]) + 1
|
||||
if attempt > int(row["max_attempts"]):
|
||||
raise TaskProtocolError("TASK_ATTEMPTS_EXHAUSTED", 409)
|
||||
expires = self.now() + timedelta(seconds=lease_seconds)
|
||||
now = iso(self.now())
|
||||
updated = connection.execute(
|
||||
"UPDATE tasks SET status='CLAIMED',claimed_by=?,lease_token_hash=?,"
|
||||
"lease_expires_at=?,attempt=?,revision=revision+1,updated_at=? "
|
||||
"WHERE task_id=? AND status='QUEUED'",
|
||||
(
|
||||
agent_id,
|
||||
self._token_hash(lease_token),
|
||||
iso(expires),
|
||||
attempt,
|
||||
now,
|
||||
task_id,
|
||||
),
|
||||
)
|
||||
if updated.rowcount != 1:
|
||||
raise TaskProtocolError("TASK_CLAIM_CONFLICT", 409)
|
||||
self._event(
|
||||
connection,
|
||||
"TASK_CLAIMED",
|
||||
task_id=task_id,
|
||||
agent_id=agent_id,
|
||||
payload={"attempt": attempt, "lease_expires_at": iso(expires)},
|
||||
)
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
connection.commit()
|
||||
return {
|
||||
**self._row_task(row, include_result=False),
|
||||
"lease_token": lease_token,
|
||||
"lease_token_returned_once": True,
|
||||
}
|
||||
|
||||
def _validate_lease(
|
||||
self,
|
||||
connection: sqlite3.Connection,
|
||||
*,
|
||||
task_id: str,
|
||||
agent_id: str,
|
||||
lease_token: str,
|
||||
allowed_states: frozenset[str] = ACTIVE_LEASE_STATES,
|
||||
) -> sqlite3.Row:
|
||||
self._expire_locked(connection)
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
if row["status"] not in allowed_states:
|
||||
raise TaskProtocolError("TASK_LEASE_NOT_ACTIVE", 409)
|
||||
if row["claimed_by"] != agent_id:
|
||||
raise TaskProtocolError("TASK_LEASE_OWNER_MISMATCH", 403)
|
||||
if not row["lease_token_hash"] or not hmac.compare_digest(
|
||||
row["lease_token_hash"], self._token_hash(lease_token)
|
||||
):
|
||||
raise TaskProtocolError("TASK_LEASE_TOKEN_INVALID", 403)
|
||||
expires = parse_time(row["lease_expires_at"])
|
||||
if expires is None or expires <= self.now():
|
||||
raise TaskProtocolError("TASK_LEASE_EXPIRED", 409)
|
||||
return row
|
||||
|
||||
def heartbeat(
|
||||
self,
|
||||
*,
|
||||
task_id: str,
|
||||
agent_id: str,
|
||||
lease_token: str,
|
||||
progress: Mapping[str, Any] | None = None,
|
||||
lease_seconds: int = 300,
|
||||
) -> dict[str, Any]:
|
||||
if not 30 <= int(lease_seconds) <= 3600:
|
||||
raise TaskProtocolError("LEASE_SECONDS_INVALID", 400)
|
||||
progress_value = _json_object(progress, "progress", max_bytes=64 * 1024)
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
self._validate_lease(
|
||||
connection,
|
||||
task_id=task_id,
|
||||
agent_id=agent_id,
|
||||
lease_token=lease_token,
|
||||
)
|
||||
expires = self.now() + timedelta(seconds=int(lease_seconds))
|
||||
now = iso(self.now())
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status='RUNNING',lease_expires_at=?,revision=revision+1,"
|
||||
"updated_at=? WHERE task_id=?",
|
||||
(iso(expires), now, task_id),
|
||||
)
|
||||
self._event(
|
||||
connection,
|
||||
"TASK_HEARTBEAT",
|
||||
task_id=task_id,
|
||||
agent_id=agent_id,
|
||||
payload={"progress": progress_value, "lease_expires_at": iso(expires)},
|
||||
)
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
connection.commit()
|
||||
return self._row_task(row, include_result=False)
|
||||
|
||||
def submit_result(
|
||||
self,
|
||||
*,
|
||||
task_id: str,
|
||||
agent_id: str,
|
||||
lease_token: str,
|
||||
idempotency_key: str,
|
||||
result: Mapping[str, Any],
|
||||
) -> dict[str, Any]:
|
||||
idempotency_key = idempotency_key.strip()
|
||||
if not idempotency_key or len(idempotency_key) > 160:
|
||||
raise TaskProtocolError("IDEMPOTENCY_KEY_INVALID", 400)
|
||||
result_value = _json_object(result, "result", max_bytes=1024 * 1024)
|
||||
result_digest = digest_json(result_value)
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
if row["result_idempotency_key"] == idempotency_key:
|
||||
if not row["result_digest"] or not hmac.compare_digest(row["result_digest"], result_digest):
|
||||
raise TaskProtocolError("IDEMPOTENCY_CONFLICT", 409)
|
||||
connection.commit()
|
||||
return {**self._row_task(row), "idempotent": True}
|
||||
self._validate_lease(
|
||||
connection,
|
||||
task_id=task_id,
|
||||
agent_id=agent_id,
|
||||
lease_token=lease_token,
|
||||
)
|
||||
now = iso(self.now())
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status='VERIFYING',result_json=?,result_digest=?,"
|
||||
"result_idempotency_key=?,lease_token_hash=NULL,lease_expires_at=NULL,"
|
||||
"revision=revision+1,updated_at=? WHERE task_id=?",
|
||||
(
|
||||
canonical_json(result_value),
|
||||
result_digest,
|
||||
idempotency_key,
|
||||
now,
|
||||
task_id,
|
||||
),
|
||||
)
|
||||
self._event(
|
||||
connection,
|
||||
"RESULT_SUBMITTED",
|
||||
task_id=task_id,
|
||||
agent_id=agent_id,
|
||||
payload={"result_digest": result_digest, "idempotency_key": idempotency_key},
|
||||
)
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
connection.commit()
|
||||
return self._row_task(row)
|
||||
|
||||
def verify_task(
|
||||
self,
|
||||
*,
|
||||
task_id: str,
|
||||
decision: str,
|
||||
verification: Mapping[str, Any] | None = None,
|
||||
) -> dict[str, Any]:
|
||||
decision = decision.strip().upper()
|
||||
if decision not in VERIFY_DECISIONS:
|
||||
raise TaskProtocolError("VERIFICATION_DECISION_INVALID", 400)
|
||||
verification_value = _json_object(verification, "verification")
|
||||
verification_value = {"decision": decision, **verification_value}
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
if row["status"] != "VERIFYING":
|
||||
raise TaskProtocolError("TASK_NOT_VERIFYING", 409)
|
||||
if decision == "SUCCEEDED":
|
||||
next_state = "SUCCEEDED"
|
||||
elif decision == "FAILED":
|
||||
next_state = "FAILED"
|
||||
elif int(row["attempt"]) >= int(row["max_attempts"]):
|
||||
next_state = "FAILED"
|
||||
verification_value["reason"] = verification_value.get(
|
||||
"reason", "MAX_ATTEMPTS_EXHAUSTED"
|
||||
)
|
||||
else:
|
||||
next_state = "QUEUED"
|
||||
now = iso(self.now())
|
||||
if next_state in {"SUCCEEDED", "FAILED"}:
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status=?,lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,verification_json=?,revision=revision+1,updated_at=? "
|
||||
"WHERE task_id=?",
|
||||
(next_state, canonical_json(verification_value), now, task_id),
|
||||
)
|
||||
else:
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status=?,claimed_by=NULL,lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,verification_json=?,revision=revision+1,updated_at=? "
|
||||
"WHERE task_id=?",
|
||||
(next_state, canonical_json(verification_value), now, task_id),
|
||||
)
|
||||
self._event(
|
||||
connection,
|
||||
"TASK_VERIFIED",
|
||||
task_id=task_id,
|
||||
agent_id=row["claimed_by"],
|
||||
payload={"decision": decision, "next_state": next_state},
|
||||
)
|
||||
updated = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
connection.commit()
|
||||
return self._row_task(updated)
|
||||
|
||||
def cancel_task(self, *, task_id: str, reason: str = "") -> dict[str, Any]:
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
if row["status"] in TERMINAL_STATES:
|
||||
connection.commit()
|
||||
return self._row_task(row)
|
||||
connection.execute(
|
||||
"UPDATE tasks SET status='CANCELLED',lease_token_hash=NULL,"
|
||||
"lease_expires_at=NULL,revision=revision+1,updated_at=? WHERE task_id=?",
|
||||
(iso(self.now()), task_id),
|
||||
)
|
||||
self._event(connection, "TASK_CANCELLED", task_id=task_id, payload={"reason": reason})
|
||||
updated = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
connection.commit()
|
||||
return self._row_task(updated)
|
||||
|
||||
def get_task(self, task_id: str) -> dict[str, Any]:
|
||||
self.expire_leases()
|
||||
with self._connect() as connection:
|
||||
row = connection.execute("SELECT * FROM tasks WHERE task_id=?", (task_id,)).fetchone()
|
||||
if row is None:
|
||||
raise TaskProtocolError("TASK_NOT_FOUND", 404)
|
||||
return self._row_task(row)
|
||||
|
||||
def get_events(self, task_id: str, *, limit: int = 200) -> list[dict[str, Any]]:
|
||||
with self._connect() as connection:
|
||||
rows = connection.execute(
|
||||
"SELECT * FROM events WHERE task_id=? ORDER BY event_id DESC LIMIT ?",
|
||||
(task_id, max(1, min(int(limit), 1000))),
|
||||
).fetchall()
|
||||
return [
|
||||
{
|
||||
"event_id": row["event_id"],
|
||||
"task_id": row["task_id"],
|
||||
"agent_id": row["agent_id"],
|
||||
"event_type": row["event_type"],
|
||||
"payload": json.loads(row["payload_json"]),
|
||||
"created_at": row["created_at"],
|
||||
}
|
||||
for row in reversed(rows)
|
||||
]
|
||||
@@ -0,0 +1,655 @@
|
||||
"""CAS-bound projection from an unprivileged planner to a host coordinator."""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
import hmac
|
||||
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
|
||||
import json
|
||||
import os
|
||||
from pathlib import Path
|
||||
import re
|
||||
import sqlite3
|
||||
from typing import Any, Callable, Mapping, Sequence
|
||||
import urllib.error
|
||||
import urllib.request
|
||||
|
||||
from aios_agent_task_initial_execution_projection import (
|
||||
InitialExecutionProjectionError,
|
||||
InitialExecutionProjectionStore,
|
||||
JOB_RE,
|
||||
KEY_RE,
|
||||
MAX_BODY,
|
||||
SHA_RE,
|
||||
TASK_RE,
|
||||
_canonical,
|
||||
_now,
|
||||
_required_authorization_id,
|
||||
_require_string,
|
||||
_sha,
|
||||
load_control_token,
|
||||
)
|
||||
from aios_online_agent_execution import (
|
||||
GovernedExecutionBinding,
|
||||
validate_execution_envelope,
|
||||
)
|
||||
|
||||
|
||||
RUNTIME_CAPABILITY = "agent.runtime.openclaw"
|
||||
COORDINATOR_CAPABILITY = "engineering.autonomous"
|
||||
COORDINATOR_CAPABILITIES = [RUNTIME_CAPABILITY, COORDINATOR_CAPABILITY]
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class SupervisorCoordinatorProjectionRequest:
|
||||
expected_status: str
|
||||
expected_revision: int
|
||||
expected_attempt: int
|
||||
expected_metadata_sha256: str
|
||||
expected_required_capabilities_sha256: str
|
||||
previous_verification_sha256: str
|
||||
expected_planner_result_sha256: str
|
||||
expected_old_execution_digest: str
|
||||
expected_new_execution_digest: str
|
||||
expected_job_id: str
|
||||
authorization_id: str
|
||||
idempotency_key: str
|
||||
reason: str
|
||||
|
||||
@classmethod
|
||||
def from_mapping(
|
||||
cls, value: Mapping[str, Any]
|
||||
) -> "SupervisorCoordinatorProjectionRequest":
|
||||
allowed = {
|
||||
"expected_status",
|
||||
"expected_revision",
|
||||
"expected_attempt",
|
||||
"expected_metadata_sha256",
|
||||
"expected_required_capabilities_sha256",
|
||||
"previous_verification_sha256",
|
||||
"expected_planner_result_sha256",
|
||||
"expected_old_execution_digest",
|
||||
"expected_new_execution_digest",
|
||||
"expected_job_id",
|
||||
"authorization_id",
|
||||
"idempotency_key",
|
||||
"reason",
|
||||
}
|
||||
if set(value) != allowed:
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_PROJECTION_FIELDS_INVALID", 400
|
||||
)
|
||||
status = _require_string(value["expected_status"], "EXPECTED_STATUS_INVALID")
|
||||
revision = value["expected_revision"]
|
||||
attempt = value["expected_attempt"]
|
||||
if (
|
||||
status != "QUEUED"
|
||||
or not isinstance(revision, int)
|
||||
or isinstance(revision, bool)
|
||||
or revision < 2
|
||||
or not isinstance(attempt, int)
|
||||
or isinstance(attempt, bool)
|
||||
or attempt < 1
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_PROJECTION_STATE_INVALID", 400
|
||||
)
|
||||
names = (
|
||||
"expected_metadata_sha256",
|
||||
"expected_required_capabilities_sha256",
|
||||
"previous_verification_sha256",
|
||||
"expected_planner_result_sha256",
|
||||
"expected_old_execution_digest",
|
||||
"expected_new_execution_digest",
|
||||
)
|
||||
digests = {
|
||||
name: _require_string(value[name], name.upper() + "_INVALID", SHA_RE)
|
||||
for name in names
|
||||
}
|
||||
job_id = _require_string(value["expected_job_id"], "JOB_ID_INVALID", JOB_RE)
|
||||
authorization_id = _require_string(
|
||||
value["authorization_id"], "AUTHORIZATION_ID_INVALID", KEY_RE
|
||||
)
|
||||
idempotency_key = _require_string(
|
||||
value["idempotency_key"], "IDEMPOTENCY_KEY_INVALID", KEY_RE
|
||||
)
|
||||
reason = _require_string(value["reason"], "PROJECTION_REASON_INVALID")
|
||||
if len(reason) > 512:
|
||||
raise InitialExecutionProjectionError("PROJECTION_REASON_INVALID", 400)
|
||||
return cls(
|
||||
status,
|
||||
revision,
|
||||
attempt,
|
||||
digests["expected_metadata_sha256"],
|
||||
digests["expected_required_capabilities_sha256"],
|
||||
digests["previous_verification_sha256"],
|
||||
digests["expected_planner_result_sha256"],
|
||||
digests["expected_old_execution_digest"],
|
||||
digests["expected_new_execution_digest"],
|
||||
job_id,
|
||||
authorization_id,
|
||||
idempotency_key,
|
||||
reason,
|
||||
)
|
||||
|
||||
def public(self) -> dict[str, Any]:
|
||||
return {
|
||||
"expected_status": self.expected_status,
|
||||
"expected_revision": self.expected_revision,
|
||||
"expected_attempt": self.expected_attempt,
|
||||
"expected_metadata_sha256": self.expected_metadata_sha256,
|
||||
"expected_required_capabilities_sha256": self.expected_required_capabilities_sha256,
|
||||
"previous_verification_sha256": self.previous_verification_sha256,
|
||||
"expected_planner_result_sha256": self.expected_planner_result_sha256,
|
||||
"expected_old_execution_digest": self.expected_old_execution_digest,
|
||||
"expected_new_execution_digest": self.expected_new_execution_digest,
|
||||
"expected_job_id": self.expected_job_id,
|
||||
"authorization_id": self.authorization_id,
|
||||
"idempotency_key": self.idempotency_key,
|
||||
"reason": self.reason,
|
||||
}
|
||||
|
||||
|
||||
class CanonicalCoordinatorExecutionResolver:
|
||||
def __init__(self, base_url: str, control_token: str) -> None:
|
||||
self.base_url = base_url.rstrip("/")
|
||||
self.control_token = control_token
|
||||
|
||||
def __call__(
|
||||
self,
|
||||
task_id: str,
|
||||
expected_job_id: str,
|
||||
runtime_input: Mapping[str, Any],
|
||||
) -> dict[str, Any]:
|
||||
body = {
|
||||
"task_id": task_id,
|
||||
"job_id": expected_job_id,
|
||||
"required_capabilities": COORDINATOR_CAPABILITIES,
|
||||
"execution": {
|
||||
"action": "agent_runtime",
|
||||
"adapter": "openclaw",
|
||||
"capability_scope": COORDINATOR_CAPABILITIES,
|
||||
"runtime_input": dict(runtime_input),
|
||||
},
|
||||
}
|
||||
request = urllib.request.Request(
|
||||
self.base_url + "/v2/execution-envelope",
|
||||
method="POST",
|
||||
data=_canonical(body),
|
||||
headers={
|
||||
"Authorization": "Bearer " + self.control_token,
|
||||
"Content-Type": "application/json",
|
||||
"Accept": "application/json",
|
||||
},
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(request, timeout=15) as response:
|
||||
raw, status = response.read(), response.status
|
||||
except urllib.error.HTTPError as exc:
|
||||
raw, status = exc.read(), exc.code
|
||||
try:
|
||||
value = json.loads(raw.decode("utf-8"))
|
||||
except (UnicodeDecodeError, json.JSONDecodeError) as exc:
|
||||
raise InitialExecutionProjectionError(
|
||||
"CANONICAL_COORDINATOR_EXECUTION_RESPONSE_INVALID", 502
|
||||
) from exc
|
||||
if not isinstance(value, Mapping):
|
||||
raise InitialExecutionProjectionError(
|
||||
"CANONICAL_COORDINATOR_EXECUTION_RESPONSE_INVALID", 502
|
||||
)
|
||||
if status != 200:
|
||||
raise InitialExecutionProjectionError(
|
||||
str(value.get("error") or "CANONICAL_COORDINATOR_EXECUTION_FAILED"),
|
||||
502 if status >= 500 else 409,
|
||||
)
|
||||
return dict(value)
|
||||
|
||||
|
||||
class SupervisorCoordinatorProjectionStore(InitialExecutionProjectionStore):
|
||||
def __init__(
|
||||
self,
|
||||
db_path: Path,
|
||||
resolver: Callable[[str, str, Mapping[str, Any]], Mapping[str, Any]],
|
||||
) -> None:
|
||||
super().__init__(db_path, resolver) # type: ignore[arg-type]
|
||||
self.coordinator_resolver = resolver
|
||||
|
||||
@staticmethod
|
||||
def _ensure_table(connection: sqlite3.Connection) -> None:
|
||||
connection.execute(
|
||||
"""
|
||||
CREATE TABLE IF NOT EXISTS task_supervisor_coordinator_projections (
|
||||
task_id TEXT NOT NULL,
|
||||
idempotency_key TEXT NOT NULL,
|
||||
request_sha256 TEXT NOT NULL,
|
||||
authorization_id TEXT NOT NULL,
|
||||
expected_job_id TEXT NOT NULL,
|
||||
planner_result_sha256 TEXT NOT NULL,
|
||||
old_execution_digest TEXT NOT NULL,
|
||||
new_execution_digest TEXT NOT NULL,
|
||||
expected_revision INTEGER NOT NULL,
|
||||
applied_revision INTEGER NOT NULL,
|
||||
attempt INTEGER NOT NULL,
|
||||
created_at TEXT NOT NULL,
|
||||
PRIMARY KEY(task_id,idempotency_key)
|
||||
)
|
||||
"""
|
||||
)
|
||||
|
||||
@classmethod
|
||||
def _planner_event_chain(
|
||||
cls, connection: sqlite3.Connection, task_id: str
|
||||
) -> None:
|
||||
table, columns = cls._event_table(connection)
|
||||
ordering = "event_id" if "event_id" in columns else "rowid"
|
||||
rows = connection.execute(
|
||||
f'SELECT event_type,payload_json FROM "{table}" '
|
||||
f'WHERE task_id=? ORDER BY {ordering}',
|
||||
(task_id,),
|
||||
).fetchall()
|
||||
events = [str(row[0]) for row in rows]
|
||||
try:
|
||||
projection = len(events) - 1 - events[::-1].index(
|
||||
"TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZED"
|
||||
)
|
||||
except ValueError as exc:
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_PROJECTION_EVENT_MISSING", 409
|
||||
) from exc
|
||||
tail = events[projection + 1 :]
|
||||
if (
|
||||
len(tail) < 4
|
||||
or tail[0] != "TASK_CLAIMED"
|
||||
or tail.count("TASK_CLAIMED") != 1
|
||||
or tail.count("RESULT_SUBMITTED") != 1
|
||||
or tail.count("TASK_VERIFIED") != 1
|
||||
or tail[-1] != "TASK_VERIFIED"
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_EVENT_CHAIN_INVALID", 409
|
||||
)
|
||||
result = tail.index("RESULT_SUBMITTED")
|
||||
if (
|
||||
not any(event == "TASK_HEARTBEAT" for event in tail[1:result])
|
||||
or any(
|
||||
event not in {
|
||||
"TASK_CLAIMED",
|
||||
"TASK_HEARTBEAT",
|
||||
"RESULT_SUBMITTED",
|
||||
"TASK_VERIFIED",
|
||||
}
|
||||
for event in tail
|
||||
)
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_EVENT_CHAIN_INVALID", 409
|
||||
)
|
||||
verification_event = json.loads(rows[-1][1])
|
||||
if (
|
||||
verification_event.get("decision") != "NEEDS_REWORK"
|
||||
or verification_event.get("next_state") != "QUEUED"
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_NEEDS_REWORK_EVENT_REQUIRED", 409
|
||||
)
|
||||
|
||||
def _load(
|
||||
self,
|
||||
connection: sqlite3.Connection,
|
||||
task_id: str,
|
||||
request: SupervisorCoordinatorProjectionRequest,
|
||||
) -> tuple[sqlite3.Row, dict[str, Any], dict[str, Any]]:
|
||||
task = connection.execute(
|
||||
"SELECT * FROM tasks WHERE task_id=?", (task_id,)
|
||||
).fetchone()
|
||||
if task is None:
|
||||
raise InitialExecutionProjectionError("TASK_NOT_FOUND", 404)
|
||||
if (
|
||||
task["status"] != request.expected_status
|
||||
or int(task["revision"]) != request.expected_revision
|
||||
or int(task["attempt"]) != request.expected_attempt
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_PROJECTION_CAS_MISMATCH", 409
|
||||
)
|
||||
if (
|
||||
task["claimed_by"] not in (None, "")
|
||||
or task["lease_token_hash"] not in (None, "")
|
||||
or task["lease_expires_at"] not in (None, "")
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_ACTIVE_LEASE_DENY", 409
|
||||
)
|
||||
metadata = self._json_value(task["metadata_json"], "TASK_METADATA_INVALID")
|
||||
capabilities = self._json_value(
|
||||
task["required_capabilities_json"], "TASK_CAPABILITIES_INVALID"
|
||||
)
|
||||
result = self._json_value(task["result_json"], "PLANNER_RESULT_INVALID")
|
||||
verification = self._json_value(
|
||||
task["verification_json"], "PLANNER_VERIFICATION_INVALID"
|
||||
)
|
||||
if (
|
||||
not isinstance(metadata, Mapping)
|
||||
or not isinstance(capabilities, list)
|
||||
or not isinstance(result, Mapping)
|
||||
or not isinstance(verification, Mapping)
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_STATE_INVALID", 409
|
||||
)
|
||||
metadata, result, verification = dict(metadata), dict(result), dict(verification)
|
||||
if _sha(metadata) != request.expected_metadata_sha256:
|
||||
raise InitialExecutionProjectionError("TASK_METADATA_DRIFT", 409)
|
||||
if _sha(capabilities) != request.expected_required_capabilities_sha256:
|
||||
raise InitialExecutionProjectionError("TASK_CAPABILITIES_DRIFT", 409)
|
||||
if _sha(result) != request.expected_planner_result_sha256:
|
||||
raise InitialExecutionProjectionError("PLANNER_RESULT_DRIFT", 409)
|
||||
if _sha(verification) != request.previous_verification_sha256:
|
||||
raise InitialExecutionProjectionError("PLANNER_VERIFICATION_DRIFT", 409)
|
||||
old = metadata.get("execution")
|
||||
if not isinstance(old, Mapping):
|
||||
raise InitialExecutionProjectionError("TASK_EXECUTION_METADATA_MISSING", 409)
|
||||
old_execution = validate_execution_envelope(dict(old))
|
||||
if (
|
||||
old_execution.get("action") != "agent_runtime"
|
||||
or old_execution.get("adapter") != "openclaw"
|
||||
or old_execution.get("capability_scope") != [RUNTIME_CAPABILITY]
|
||||
or old_execution.get("execution_digest")
|
||||
!= request.expected_old_execution_digest
|
||||
or capabilities != [RUNTIME_CAPABILITY]
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_EXECUTION_CONTRACT_INVALID", 409
|
||||
)
|
||||
expected_result = {
|
||||
"task_id": task_id,
|
||||
"job_id": request.expected_job_id,
|
||||
"status": "SUCCEEDED",
|
||||
"execution_mode": "agent_runtime",
|
||||
"agent_runtime": "openclaw",
|
||||
"execution_digest": request.expected_old_execution_digest,
|
||||
"control_plane_access_granted_to_runtime": False,
|
||||
"credential_material_exposed": False,
|
||||
"secret_material_exposed": False,
|
||||
"changed_files": [],
|
||||
"writer_lease": "NONE",
|
||||
"direct_main_write": False,
|
||||
"production_write": False,
|
||||
"force_push": False,
|
||||
"protected_ref_bypass": False,
|
||||
}
|
||||
if any(result.get(name) != expected for name, expected in expected_result.items()):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_RESULT_CONTRACT_INVALID", 409
|
||||
)
|
||||
if verification.get("decision") != "NEEDS_REWORK":
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_NEEDS_REWORK_REQUIRED", 409
|
||||
)
|
||||
self._planner_event_chain(connection, task_id)
|
||||
return task, metadata, result
|
||||
|
||||
def project(
|
||||
self,
|
||||
task_id: str,
|
||||
request: SupervisorCoordinatorProjectionRequest,
|
||||
) -> dict[str, Any]:
|
||||
task_id = _require_string(task_id, "TASK_ID_INVALID", TASK_RE)
|
||||
if request.authorization_id != _required_authorization_id():
|
||||
raise InitialExecutionProjectionError("AUTHORIZATION_ID_MISMATCH", 403)
|
||||
request_sha = _sha(request.public())
|
||||
with self._connect() as connection:
|
||||
self._ensure_table(connection)
|
||||
replay = connection.execute(
|
||||
"SELECT * FROM task_supervisor_coordinator_projections "
|
||||
"WHERE task_id=? AND idempotency_key=?",
|
||||
(task_id, request.idempotency_key),
|
||||
).fetchone()
|
||||
if replay is not None:
|
||||
if replay["request_sha256"] != request_sha:
|
||||
raise InitialExecutionProjectionError("IDEMPOTENCY_CONFLICT", 409)
|
||||
return {
|
||||
"task_id": task_id,
|
||||
"status": "QUEUED",
|
||||
"attempt": replay["attempt"],
|
||||
"revision": replay["applied_revision"],
|
||||
"execution_digest": replay["new_execution_digest"],
|
||||
"required_capabilities": COORDINATOR_CAPABILITIES,
|
||||
"idempotent": True,
|
||||
}
|
||||
with self._connect() as connection:
|
||||
task, _, result = self._load(connection, task_id, request)
|
||||
runtime_input = self._runtime_input(task)
|
||||
resolved = dict(
|
||||
self.coordinator_resolver(task_id, request.expected_job_id, runtime_input)
|
||||
)
|
||||
binding = GovernedExecutionBinding.from_mapping(resolved)
|
||||
if binding.task_id != task_id or binding.job_id != request.expected_job_id:
|
||||
raise InitialExecutionProjectionError("GOVERNED_BINDING_CONFLICT", 409)
|
||||
result_binding = {
|
||||
name: result.get(name)
|
||||
for name in (
|
||||
"task_id",
|
||||
"job_id",
|
||||
"repository",
|
||||
"base_oid",
|
||||
"expected_branch_oid",
|
||||
"target_branch",
|
||||
"generation",
|
||||
)
|
||||
}
|
||||
expected_binding = {
|
||||
"task_id": binding.task_id,
|
||||
"job_id": binding.job_id,
|
||||
"repository": binding.repository,
|
||||
"base_oid": binding.base_oid,
|
||||
"expected_branch_oid": binding.expected_branch_oid,
|
||||
"target_branch": binding.target_branch,
|
||||
"generation": binding.generation,
|
||||
}
|
||||
if result_binding != expected_binding:
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_PLANNER_BINDING_CONFLICT", 409
|
||||
)
|
||||
execution = validate_execution_envelope(resolved, binding=binding)
|
||||
if (
|
||||
execution.get("action") != "agent_runtime"
|
||||
or execution.get("adapter") != "openclaw"
|
||||
or execution.get("capability_scope") != COORDINATOR_CAPABILITIES
|
||||
or execution.get("change_set") != []
|
||||
or execution.get("verification") != []
|
||||
or execution.get("governance", {}).get("writer_lease") is not False
|
||||
or execution.get("execution_digest") != request.expected_new_execution_digest
|
||||
):
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_ENVELOPE_INVALID", 409
|
||||
)
|
||||
with self._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
self._ensure_table(connection)
|
||||
task, metadata, _ = self._load(connection, task_id, request)
|
||||
new_metadata = dict(metadata)
|
||||
new_metadata["execution"] = execution
|
||||
new_metadata["execution_projection"] = {
|
||||
"schema": "aios.online-agent.task-projection.v1",
|
||||
"task_id": task_id,
|
||||
"job_id": request.expected_job_id,
|
||||
"execution_digest": execution["execution_digest"],
|
||||
}
|
||||
applied_revision = request.expected_revision + 1
|
||||
cursor = connection.execute(
|
||||
"""
|
||||
UPDATE tasks SET metadata_json=?,required_capabilities_json=?,
|
||||
revision=?,updated_at=?
|
||||
WHERE task_id=? AND status=? AND revision=? AND attempt=?
|
||||
AND claimed_by IS NULL AND lease_token_hash IS NULL
|
||||
AND lease_expires_at IS NULL AND metadata_json=?
|
||||
AND required_capabilities_json=? AND result_json=?
|
||||
AND verification_json=?
|
||||
""",
|
||||
(
|
||||
json.dumps(new_metadata, sort_keys=True, separators=(",", ":")),
|
||||
json.dumps(COORDINATOR_CAPABILITIES, separators=(",", ":")),
|
||||
applied_revision,
|
||||
_now(),
|
||||
task_id,
|
||||
request.expected_status,
|
||||
request.expected_revision,
|
||||
request.expected_attempt,
|
||||
task["metadata_json"],
|
||||
task["required_capabilities_json"],
|
||||
task["result_json"],
|
||||
task["verification_json"],
|
||||
),
|
||||
)
|
||||
if cursor.rowcount != 1:
|
||||
raise InitialExecutionProjectionError(
|
||||
"SUPERVISOR_COORDINATOR_PROJECTION_CAS_MISMATCH", 409
|
||||
)
|
||||
event = {
|
||||
"schema": "aios.supervisor-coordinator-projection.event.v1",
|
||||
"task_id": task_id,
|
||||
"job_id": request.expected_job_id,
|
||||
"authorization_id": request.authorization_id,
|
||||
"idempotency_key": request.idempotency_key,
|
||||
"reason": request.reason,
|
||||
"old_execution_digest": request.expected_old_execution_digest,
|
||||
"new_execution_digest": execution["execution_digest"],
|
||||
"required_capabilities": COORDINATOR_CAPABILITIES,
|
||||
"previous_revision": request.expected_revision,
|
||||
"applied_revision": applied_revision,
|
||||
"attempt": request.expected_attempt,
|
||||
"request_sha256": request_sha,
|
||||
"credential_forwarded_to_planner": False,
|
||||
"writer_lease": False,
|
||||
"main_write": False,
|
||||
"production_write": False,
|
||||
}
|
||||
table, columns = self._event_table(connection)
|
||||
values = {
|
||||
"task_id": task_id,
|
||||
"event_type": "TASK_SUPERVISOR_COORDINATOR_PROJECTION_AUTHORIZED",
|
||||
"payload_json": json.dumps(event, sort_keys=True, separators=(",", ":")),
|
||||
"created_at": _now(),
|
||||
"agent_id": "owner-control",
|
||||
}
|
||||
names = [name for name in values if name in columns]
|
||||
connection.execute(
|
||||
f'INSERT INTO "{table}" ({",".join(chr(34)+name+chr(34) for name in names)}) '
|
||||
f'VALUES ({",".join("?" for _ in names)})',
|
||||
tuple(values[name] for name in names),
|
||||
)
|
||||
connection.execute(
|
||||
"""
|
||||
INSERT INTO task_supervisor_coordinator_projections(
|
||||
task_id,idempotency_key,request_sha256,authorization_id,
|
||||
expected_job_id,planner_result_sha256,old_execution_digest,
|
||||
new_execution_digest,expected_revision,applied_revision,
|
||||
attempt,created_at
|
||||
) VALUES (?,?,?,?,?,?,?,?,?,?,?,?)
|
||||
""",
|
||||
(
|
||||
task_id,
|
||||
request.idempotency_key,
|
||||
request_sha,
|
||||
request.authorization_id,
|
||||
request.expected_job_id,
|
||||
request.expected_planner_result_sha256,
|
||||
request.expected_old_execution_digest,
|
||||
execution["execution_digest"],
|
||||
request.expected_revision,
|
||||
applied_revision,
|
||||
request.expected_attempt,
|
||||
_now(),
|
||||
),
|
||||
)
|
||||
connection.commit()
|
||||
return {
|
||||
"task_id": task_id,
|
||||
"status": "QUEUED",
|
||||
"attempt": request.expected_attempt,
|
||||
"revision": applied_revision,
|
||||
"execution_digest": execution["execution_digest"],
|
||||
"required_capabilities": COORDINATOR_CAPABILITIES,
|
||||
"credential_forwarded_to_planner": False,
|
||||
"idempotent": False,
|
||||
}
|
||||
|
||||
|
||||
class Handler(BaseHTTPRequestHandler):
|
||||
store: SupervisorCoordinatorProjectionStore | None = None
|
||||
control_token: str = ""
|
||||
|
||||
def log_message(self, *_args: Any) -> None:
|
||||
return
|
||||
|
||||
def send_json(self, value: Mapping[str, Any], status: int) -> None:
|
||||
raw = _canonical(dict(value))
|
||||
self.send_response(status)
|
||||
self.send_header("Content-Type", "application/json")
|
||||
self.send_header("Content-Length", str(len(raw)))
|
||||
self.end_headers()
|
||||
self.wfile.write(raw)
|
||||
|
||||
def do_GET(self) -> None:
|
||||
if self.path == "/health":
|
||||
self.send_json(
|
||||
{"status": "ok", "service": "aios-supervisor-coordinator-projection"},
|
||||
200,
|
||||
)
|
||||
else:
|
||||
self.send_json({"error": "NOT_FOUND"}, 404)
|
||||
|
||||
def do_POST(self) -> None:
|
||||
match = re.fullmatch(
|
||||
r"/v1/tasks/([^/]+)/supervisor-coordinator-projection", self.path
|
||||
)
|
||||
header = self.headers.get("Authorization", "")
|
||||
if match is None:
|
||||
self.send_json({"error": "NOT_FOUND"}, 404)
|
||||
return
|
||||
if (
|
||||
not header.startswith("Bearer ")
|
||||
or not hmac.compare_digest(header[7:], self.control_token)
|
||||
):
|
||||
self.send_json({"error": "UNAUTHORIZED"}, 401)
|
||||
return
|
||||
try:
|
||||
length = int(self.headers.get("Content-Length", "0"))
|
||||
if length < 1 or length > MAX_BODY:
|
||||
raise InitialExecutionProjectionError("REQUEST_SIZE_INVALID", 413)
|
||||
value = json.loads(self.rfile.read(length).decode("utf-8"))
|
||||
if not isinstance(value, Mapping) or self.store is None:
|
||||
raise InitialExecutionProjectionError("REQUEST_INVALID", 400)
|
||||
request = SupervisorCoordinatorProjectionRequest.from_mapping(value)
|
||||
result = self.store.project(match.group(1), request)
|
||||
self.send_json(result, 200 if result["idempotent"] else 201)
|
||||
except InitialExecutionProjectionError as exc:
|
||||
self.send_json({"error": exc.code}, exc.status)
|
||||
except (ValueError, UnicodeDecodeError, json.JSONDecodeError):
|
||||
self.send_json({"error": "REQUEST_INVALID"}, 400)
|
||||
|
||||
|
||||
def serve() -> None:
|
||||
credential = Path(
|
||||
os.environ.get(
|
||||
"AIOS_AGENT_TASK_CONTROL_TOKEN_FILE",
|
||||
str(
|
||||
Path(os.environ.get("CREDENTIALS_DIRECTORY", "/run/credentials"))
|
||||
/ "agent_task_control_token"
|
||||
),
|
||||
)
|
||||
)
|
||||
token = load_control_token(credential)
|
||||
bridge = os.environ.get("AIOS_GOVERNED_WRITE_API_URL", "http://127.0.0.1:18892")
|
||||
db = Path(os.environ.get("AIOS_AGENT_TASK_DB", "/var/lib/aios-agent-task/state.sqlite3"))
|
||||
Handler.control_token = token
|
||||
Handler.store = SupervisorCoordinatorProjectionStore(
|
||||
db, CanonicalCoordinatorExecutionResolver(bridge, token)
|
||||
)
|
||||
host = os.environ.get("AIOS_SUPERVISOR_COORDINATOR_PROJECTION_HOST", "127.0.0.1")
|
||||
port = int(os.environ.get("AIOS_SUPERVISOR_COORDINATOR_PROJECTION_PORT", "18944"))
|
||||
ThreadingHTTPServer((host, port), Handler).serve_forever()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
serve()
|
||||
@@ -0,0 +1,288 @@
|
||||
"""Unified autonomous control-plane primitives.
|
||||
|
||||
This is deliberately a library, not another daemon. Task API, workers and host
|
||||
recovery code consume the same Admission -> Invariant -> Repair semantics.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import asdict, dataclass
|
||||
from enum import Enum
|
||||
import re
|
||||
from typing import Callable, Iterable, Mapping, Sequence
|
||||
|
||||
from aios_capability_contract import (
|
||||
capability_contract,
|
||||
has_capabilities,
|
||||
normalize_required_capabilities,
|
||||
normalize_worker_capabilities,
|
||||
registered_capability,
|
||||
validate_capability_contract_registry,
|
||||
)
|
||||
|
||||
|
||||
class FailureClass(str, Enum):
|
||||
CAPABILITY_CONTRACT_ERROR = "CAPABILITY_CONTRACT_ERROR"
|
||||
CREDENTIAL_PROJECTION_ERROR = "CREDENTIAL_PROJECTION_ERROR"
|
||||
DEPENDENCY_UNAVAILABLE = "DEPENDENCY_UNAVAILABLE"
|
||||
POLICY_DENIED = "POLICY_DENIED"
|
||||
RESOURCE_UNAVAILABLE = "RESOURCE_UNAVAILABLE"
|
||||
WORKER_CRASH = "WORKER_CRASH"
|
||||
LEASE_LOST = "LEASE_LOST"
|
||||
BUSINESS_PERMANENT = "BUSINESS_PERMANENT"
|
||||
UNKNOWN = "UNKNOWN"
|
||||
|
||||
|
||||
class RepairLevel(int, Enum):
|
||||
AUTO_LOCAL = 0
|
||||
AUTO_REPAIR_TASK = 1
|
||||
OWNER_APPROVAL = 2
|
||||
MANUAL = 3
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class InvariantResult:
|
||||
name: str
|
||||
passed: bool
|
||||
detail: str = ""
|
||||
|
||||
def as_dict(self) -> dict[str, object]:
|
||||
return asdict(self)
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class AdmissionDecision:
|
||||
admitted: bool
|
||||
state: str
|
||||
invariants: tuple[InvariantResult, ...]
|
||||
failure_class: FailureClass | None = None
|
||||
repair_level: RepairLevel | None = None
|
||||
repair_action: str | None = None
|
||||
|
||||
def as_dict(self) -> dict[str, object]:
|
||||
return {
|
||||
"schema": "aios.runtime-admission.decision.v1",
|
||||
"admitted": self.admitted,
|
||||
"state": self.state,
|
||||
"invariants": [item.as_dict() for item in self.invariants],
|
||||
"failure_class": None if self.failure_class is None else self.failure_class.value,
|
||||
"repair_level": None if self.repair_level is None else int(self.repair_level),
|
||||
"repair_action": self.repair_action,
|
||||
}
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class RepairDecision:
|
||||
failure_class: FailureClass
|
||||
level: RepairLevel
|
||||
action: str
|
||||
consumes_business_attempt: bool
|
||||
|
||||
def as_dict(self) -> dict[str, object]:
|
||||
return {
|
||||
"failure_class": self.failure_class.value,
|
||||
"level": int(self.level),
|
||||
"action": self.action,
|
||||
"consumes_business_attempt": self.consumes_business_attempt,
|
||||
}
|
||||
|
||||
|
||||
def classify_failure(error: object, *, phase: str = "") -> FailureClass:
|
||||
text = (str(error) + " " + str(phase)).lower()
|
||||
if any(token in text for token in (
|
||||
"credential", "permission denied", "token registry", "secret", "unauthorized"
|
||||
)):
|
||||
return FailureClass.CREDENTIAL_PROJECTION_ERROR
|
||||
if any(token in text for token in (
|
||||
"transport", "unavailable", "timeout", "connection", "bad gateway", "http_5"
|
||||
)):
|
||||
return FailureClass.DEPENDENCY_UNAVAILABLE
|
||||
if any(token in text for token in (
|
||||
"capability", "eligible", "matcher", "contract"
|
||||
)):
|
||||
return FailureClass.CAPABILITY_CONTRACT_ERROR
|
||||
if any(token in text for token in (
|
||||
"policy", "forbidden", "approval required", "not authorized"
|
||||
)):
|
||||
return FailureClass.POLICY_DENIED
|
||||
if any(token in text for token in ("lease lost", "lease_expired", "heartbeat_failed")):
|
||||
return FailureClass.LEASE_LOST
|
||||
if any(token in text for token in ("out of memory", "disk full", "resource")):
|
||||
return FailureClass.RESOURCE_UNAVAILABLE
|
||||
if any(token in text for token in ("worker crash", "process exited", "signal")):
|
||||
return FailureClass.WORKER_CRASH
|
||||
if any(token in text for token in ("invalid release", "business permanent")):
|
||||
return FailureClass.BUSINESS_PERMANENT
|
||||
return FailureClass.UNKNOWN
|
||||
|
||||
|
||||
def repair_for_failure(failure: FailureClass) -> RepairDecision:
|
||||
table = {
|
||||
FailureClass.CAPABILITY_CONTRACT_ERROR: RepairDecision(
|
||||
failure, RepairLevel.AUTO_REPAIR_TASK,
|
||||
"create_same_lineage_capability_contract_repair", False,
|
||||
),
|
||||
FailureClass.CREDENTIAL_PROJECTION_ERROR: RepairDecision(
|
||||
failure, RepairLevel.AUTO_REPAIR_TASK,
|
||||
"restore_service_private_credential_projection", False,
|
||||
),
|
||||
FailureClass.DEPENDENCY_UNAVAILABLE: RepairDecision(
|
||||
failure, RepairLevel.AUTO_LOCAL,
|
||||
"retry_dependency_health_and_resume", False,
|
||||
),
|
||||
FailureClass.RESOURCE_UNAVAILABLE: RepairDecision(
|
||||
failure, RepairLevel.AUTO_LOCAL,
|
||||
"heal_runtime_resource_and_resume", False,
|
||||
),
|
||||
FailureClass.WORKER_CRASH: RepairDecision(
|
||||
failure, RepairLevel.AUTO_LOCAL,
|
||||
"restart_worker_and_resume_same_lineage", False,
|
||||
),
|
||||
FailureClass.LEASE_LOST: RepairDecision(
|
||||
failure, RepairLevel.AUTO_LOCAL,
|
||||
"recover_lease_and_resume_same_lineage", False,
|
||||
),
|
||||
FailureClass.POLICY_DENIED: RepairDecision(
|
||||
failure, RepairLevel.OWNER_APPROVAL,
|
||||
"request_bound_owner_authorization", False,
|
||||
),
|
||||
FailureClass.BUSINESS_PERMANENT: RepairDecision(
|
||||
failure, RepairLevel.MANUAL,
|
||||
"terminal_business_failure", True,
|
||||
),
|
||||
FailureClass.UNKNOWN: RepairDecision(
|
||||
failure, RepairLevel.AUTO_REPAIR_TASK,
|
||||
"create_diagnostic_repair_task", False,
|
||||
),
|
||||
}
|
||||
return table[failure]
|
||||
|
||||
|
||||
def _probe(
|
||||
name: str,
|
||||
probes: Mapping[str, Callable[[], bool]],
|
||||
) -> InvariantResult:
|
||||
fn = probes.get(name)
|
||||
if fn is None:
|
||||
return InvariantResult(name, False, "PROBE_MISSING")
|
||||
try:
|
||||
passed = bool(fn())
|
||||
except Exception as exc:
|
||||
return InvariantResult(name, False, "PROBE_ERROR:" + type(exc).__name__)
|
||||
return InvariantResult(name, passed, "PASS" if passed else "NOT_READY")
|
||||
|
||||
|
||||
def runtime_admission(
|
||||
*,
|
||||
worker_capabilities: Iterable[object],
|
||||
required_capabilities: Iterable[object],
|
||||
requested_operations: Sequence[str] = (),
|
||||
probes: Mapping[str, Callable[[], bool]] | None = None,
|
||||
) -> AdmissionDecision:
|
||||
"""Admission is evaluated before claim; failures do not consume attempts."""
|
||||
|
||||
worker_values = tuple(worker_capabilities)
|
||||
required_values = tuple(required_capabilities)
|
||||
probe_map = probes or {}
|
||||
invariants: list[InvariantResult] = []
|
||||
|
||||
registry = validate_capability_contract_registry()
|
||||
invariants.append(InvariantResult(
|
||||
"capability_contract_registry",
|
||||
bool(registry.get("valid")),
|
||||
"PASS" if registry.get("valid") else "REGISTRY_INVALID",
|
||||
))
|
||||
|
||||
task_known = all(registered_capability(item) for item in required_values)
|
||||
worker_known = all(registered_capability(item, worker=True) for item in worker_values)
|
||||
invariants.extend((
|
||||
InvariantResult("task_capabilities_registered", task_known, "PASS" if task_known else "UNKNOWN_TASK_CAPABILITY"),
|
||||
InvariantResult("worker_capabilities_registered", worker_known, "PASS" if worker_known else "UNKNOWN_WORKER_CAPABILITY"),
|
||||
))
|
||||
|
||||
capability_match = (
|
||||
task_known
|
||||
and worker_known
|
||||
and has_capabilities(worker_values, required_values)
|
||||
)
|
||||
invariants.append(InvariantResult(
|
||||
"capability_match",
|
||||
capability_match,
|
||||
"PASS" if capability_match else "WORKER_NOT_ELIGIBLE",
|
||||
))
|
||||
|
||||
normalized = normalize_required_capabilities(required_values)
|
||||
contracts = [
|
||||
capability_contract(item)
|
||||
for item in normalized
|
||||
if not str(item).startswith("exact:")
|
||||
]
|
||||
contracts = [item for item in contracts if item is not None]
|
||||
|
||||
forbidden: set[str] = set()
|
||||
dependencies: set[str] = set()
|
||||
health: set[str] = set()
|
||||
for contract in contracts:
|
||||
forbidden.update(contract.forbidden)
|
||||
dependencies.update(contract.dependencies)
|
||||
health.update(contract.health_requirements)
|
||||
|
||||
requested = {str(item).strip().lower() for item in requested_operations if str(item).strip()}
|
||||
forbidden_hit = sorted(requested & {item.lower() for item in forbidden})
|
||||
invariants.append(InvariantResult(
|
||||
"forbidden_operations_absent",
|
||||
not forbidden_hit,
|
||||
"PASS" if not forbidden_hit else "FORBIDDEN:" + ",".join(forbidden_hit),
|
||||
))
|
||||
|
||||
for name in sorted(dependencies | health):
|
||||
invariants.append(_probe(name, probe_map))
|
||||
|
||||
failed = [item for item in invariants if not item.passed]
|
||||
if not failed:
|
||||
return AdmissionDecision(
|
||||
True,
|
||||
"EXECUTION_READY",
|
||||
tuple(invariants),
|
||||
)
|
||||
|
||||
details = " ".join(item.name + ":" + item.detail for item in failed)
|
||||
failure = classify_failure(details, phase="admission")
|
||||
if any(item.name == "forbidden_operations_absent" and not item.passed for item in failed):
|
||||
failure = FailureClass.POLICY_DENIED
|
||||
elif any("credential" in item.name for item in failed):
|
||||
failure = FailureClass.CREDENTIAL_PROJECTION_ERROR
|
||||
elif any(item.name in {"capability_contract_registry", "task_capabilities_registered", "worker_capabilities_registered", "capability_match"} for item in failed):
|
||||
failure = FailureClass.CAPABILITY_CONTRACT_ERROR
|
||||
elif any(item.name in dependencies or item.name in health for item in failed):
|
||||
failure = FailureClass.DEPENDENCY_UNAVAILABLE
|
||||
|
||||
repair = repair_for_failure(failure)
|
||||
return AdmissionDecision(
|
||||
False,
|
||||
"EXECUTION_READY_FAILED",
|
||||
tuple(invariants),
|
||||
failure,
|
||||
repair.level,
|
||||
repair.action,
|
||||
)
|
||||
|
||||
|
||||
def system_invariant_snapshot(
|
||||
*,
|
||||
worker_capabilities: Iterable[object],
|
||||
required_capabilities: Iterable[object],
|
||||
requested_operations: Sequence[str] = (),
|
||||
probes: Mapping[str, Callable[[], bool]] | None = None,
|
||||
) -> dict[str, object]:
|
||||
decision = runtime_admission(
|
||||
worker_capabilities=worker_capabilities,
|
||||
required_capabilities=required_capabilities,
|
||||
requested_operations=requested_operations,
|
||||
probes=probes,
|
||||
)
|
||||
return {
|
||||
"schema": "aios.system-invariants.v1",
|
||||
"status": "HEALTHY" if decision.admitted else "DEGRADED",
|
||||
"execution_ready": decision.admitted,
|
||||
"admission": decision.as_dict(),
|
||||
}
|
||||
@@ -0,0 +1,142 @@
|
||||
"""Private Unix-socket caller projection and rehydration client.
|
||||
|
||||
The broker response body is deliberately public metadata only. The short-lived
|
||||
credential projection is delivered in a private response header, consumed
|
||||
inside the model API process, and never returned by an HTTP API or persisted.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
import hashlib
|
||||
import http.client
|
||||
import json
|
||||
from pathlib import Path
|
||||
import socket
|
||||
from typing import Any, Mapping
|
||||
|
||||
CALLER = "chatgpt-governed-implementer-singlehost"
|
||||
SCOPE = "release:candidate"
|
||||
GOVERNED_TASK_SCOPE = "governed-task:project"
|
||||
PROJECTED_SCOPES = frozenset({GOVERNED_TASK_SCOPE, SCOPE})
|
||||
CREDENTIAL_HEADER = "X-AIOS-Caller-Projection-Credential"
|
||||
|
||||
|
||||
class CallerProjectionError(RuntimeError):
|
||||
def __init__(self, code: str, status: int = 503) -> None:
|
||||
super().__init__(code)
|
||||
self.code = code
|
||||
self.status = status
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class ProjectionResult:
|
||||
public: dict[str, Any]
|
||||
credential_loaded: bool
|
||||
|
||||
|
||||
class _UnixHTTPConnection(http.client.HTTPConnection):
|
||||
def __init__(self, socket_path: Path, timeout: float) -> None:
|
||||
super().__init__("localhost", timeout=timeout)
|
||||
self.socket_path = socket_path
|
||||
|
||||
def connect(self) -> None:
|
||||
sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM)
|
||||
sock.settimeout(self.timeout)
|
||||
try:
|
||||
sock.connect(str(self.socket_path))
|
||||
except Exception:
|
||||
sock.close()
|
||||
raise
|
||||
self.sock = sock
|
||||
|
||||
|
||||
class CallerProjectionClient:
|
||||
def __init__(self, socket_path: Path, *, timeout: float = 2.0) -> None:
|
||||
self.socket_path = Path(socket_path)
|
||||
self.timeout = timeout
|
||||
|
||||
@staticmethod
|
||||
def _idempotency_key(caller: str, scope: str) -> str:
|
||||
digest = hashlib.sha256((caller + "\0" + scope).encode()).hexdigest()
|
||||
return "caller-projection-" + digest[:32] + "-v1"
|
||||
|
||||
def rehydrate(
|
||||
self,
|
||||
caller: str,
|
||||
scope: str,
|
||||
request: Mapping[str, Any] | None = None,
|
||||
) -> ProjectionResult:
|
||||
if caller != CALLER:
|
||||
raise CallerProjectionError("CALLER_MISMATCH", 403)
|
||||
if scope not in PROJECTED_SCOPES:
|
||||
raise CallerProjectionError("SCOPE_MISMATCH", 403)
|
||||
body = {
|
||||
"caller": caller,
|
||||
"scope": scope,
|
||||
"idempotency_key": self._idempotency_key(caller, scope),
|
||||
"request": dict(request or {}),
|
||||
}
|
||||
connection = _UnixHTTPConnection(self.socket_path, self.timeout)
|
||||
try:
|
||||
connection.request(
|
||||
"POST",
|
||||
"/v1/project-caller",
|
||||
body=json.dumps(body, sort_keys=True, separators=(",", ":")).encode(),
|
||||
headers={
|
||||
"Content-Type": "application/json",
|
||||
"Accept": "application/json",
|
||||
"X-AIOS-Account": caller,
|
||||
},
|
||||
)
|
||||
response = connection.getresponse()
|
||||
status = response.status
|
||||
raw = response.read()
|
||||
credential = response.getheader(CREDENTIAL_HEADER, "")
|
||||
except (OSError, TimeoutError, http.client.HTTPException) as exc:
|
||||
raise CallerProjectionError("PROJECTION_BROKER_UNAVAILABLE", 503) from exc
|
||||
finally:
|
||||
connection.close()
|
||||
|
||||
try:
|
||||
value = json.loads(raw.decode("utf-8"))
|
||||
except Exception as exc:
|
||||
raise CallerProjectionError("PROJECTION_RESPONSE_INVALID", 503) from exc
|
||||
if not isinstance(value, Mapping):
|
||||
raise CallerProjectionError("PROJECTION_RESPONSE_INVALID", 503)
|
||||
if status != 200:
|
||||
code = str(value.get("error") or "PROJECTION_DENIED")
|
||||
raise CallerProjectionError(code, 403 if status < 500 else 503)
|
||||
if any(
|
||||
str(key).lower() in {
|
||||
"credential",
|
||||
"credential_material",
|
||||
"token",
|
||||
"bearer",
|
||||
"secret",
|
||||
"password",
|
||||
}
|
||||
for key in value
|
||||
):
|
||||
raise CallerProjectionError("PROJECTION_SECRET_FIELD_REJECTED", 503)
|
||||
if value.get("caller") != CALLER or value.get("scope") != scope:
|
||||
raise CallerProjectionError("PROJECTION_BINDING_INVALID", 503)
|
||||
if value.get("projection_state") != "ACKED":
|
||||
raise CallerProjectionError("PROJECTION_STATE_INVALID", 503)
|
||||
if value.get("credential_state") != "REHYDRATED":
|
||||
raise CallerProjectionError("CREDENTIAL_REHYDRATION_INVALID", 503)
|
||||
if value.get("rehydration_state") not in {"GENERATED", "RECOVERED"}:
|
||||
raise CallerProjectionError("REHYDRATION_STATE_INVALID", 503)
|
||||
effective_scopes = value.get("effective_scopes")
|
||||
if (
|
||||
not isinstance(effective_scopes, list)
|
||||
or len(effective_scopes) != len(set(effective_scopes))
|
||||
or not all(isinstance(item, str) for item in effective_scopes)
|
||||
or scope not in effective_scopes
|
||||
or not set(effective_scopes).issubset(PROJECTED_SCOPES)
|
||||
):
|
||||
raise CallerProjectionError("PROJECTION_SCOPE_INVALID", 503)
|
||||
if value.get("secret_exposed") is not False:
|
||||
raise CallerProjectionError("PROJECTION_SECRET_EXPOSURE", 503)
|
||||
if not isinstance(credential, str) or len(credential) < 32:
|
||||
raise CallerProjectionError("CREDENTIAL_REHYDRATION_MISSING", 503)
|
||||
return ProjectionResult(public=dict(value), credential_loaded=True)
|
||||
@@ -3,6 +3,8 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from collections.abc import Iterable
|
||||
from dataclasses import asdict, dataclass
|
||||
from typing import Any
|
||||
|
||||
|
||||
_TASK_ALIASES: dict[str, frozenset[str]] = {
|
||||
@@ -37,6 +39,8 @@ _TASK_ALIASES: dict[str, frozenset[str]] = {
|
||||
"gpu.train": frozenset({"legacy.gpu.train"}),
|
||||
"agent.runtime.hermes": frozenset({"agent.runtime.hermes"}),
|
||||
"agent.runtime.openclaw": frozenset({"agent.runtime.openclaw"}),
|
||||
"engineering.autonomous": frozenset({"engineering.autonomous"}),
|
||||
"release:candidate": frozenset({"release:candidate"}),
|
||||
}
|
||||
|
||||
_WORKER_ALIASES: dict[str, frozenset[str]] = {
|
||||
@@ -71,8 +75,178 @@ _WORKER_ALIASES: dict[str, frozenset[str]] = {
|
||||
"gpu.train": frozenset({"legacy.gpu.train"}),
|
||||
"agent.runtime.hermes": frozenset({"agent.runtime.hermes"}),
|
||||
"agent.runtime.openclaw": frozenset({"agent.runtime.openclaw"}),
|
||||
"engineering.autonomous": frozenset({"engineering.autonomous"}),
|
||||
"release:candidate": frozenset({"release:candidate"}),
|
||||
}
|
||||
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class CapabilityContract:
|
||||
"""Versioned runtime contract attached to one canonical semantic capability."""
|
||||
|
||||
name: str
|
||||
min_version: int = 1
|
||||
max_version: int = 1
|
||||
permissions: tuple[str, ...] = ()
|
||||
forbidden: tuple[str, ...] = ()
|
||||
dependencies: tuple[str, ...] = ()
|
||||
health_requirements: tuple[str, ...] = ()
|
||||
|
||||
def as_dict(self) -> dict[str, Any]:
|
||||
return asdict(self)
|
||||
|
||||
|
||||
# This registry is the semantic source of truth. Alias tables below are retained
|
||||
# as compatibility adapters; startup/tests validate every alias target against
|
||||
# this registry so a name cannot be accepted in one layer and be unknown in
|
||||
# another.
|
||||
_CAPABILITY_CONTRACTS: dict[str, CapabilityContract] = {
|
||||
"repo.read": CapabilityContract("repo.read", permissions=("repository_read",)),
|
||||
"test.execute": CapabilityContract("test.execute", permissions=("test_execution",)),
|
||||
"governed.git.write.branch": CapabilityContract(
|
||||
"governed.git.write.branch",
|
||||
permissions=("governed_branch_write",),
|
||||
dependencies=("writer_lease", "governed_gitops"),
|
||||
),
|
||||
"governed.git.write.file": CapabilityContract(
|
||||
"governed.git.write.file",
|
||||
permissions=("governed_file_write",),
|
||||
dependencies=("writer_lease", "governed_gitops"),
|
||||
),
|
||||
"governed.git.write": CapabilityContract(
|
||||
"governed.git.write",
|
||||
permissions=("governed_git_write",),
|
||||
dependencies=("writer_lease", "governed_gitops"),
|
||||
),
|
||||
"runtime.shell": CapabilityContract("runtime.shell", permissions=("runtime_shell",)),
|
||||
"runtime.git": CapabilityContract("runtime.git", permissions=("runtime_git",)),
|
||||
"runtime.python": CapabilityContract("runtime.python", permissions=("runtime_python",)),
|
||||
"legacy.code.read": CapabilityContract("legacy.code.read"),
|
||||
"legacy.code.review": CapabilityContract("legacy.code.review"),
|
||||
"legacy.code.write": CapabilityContract("legacy.code.write"),
|
||||
"legacy.code.write.patch": CapabilityContract("legacy.code.write.patch"),
|
||||
"legacy.test.python": CapabilityContract("legacy.test.python"),
|
||||
"legacy.test.shell.allowlisted": CapabilityContract("legacy.test.shell.allowlisted"),
|
||||
"legacy.git.commit": CapabilityContract("legacy.git.commit"),
|
||||
"legacy.git.diff.read": CapabilityContract("legacy.git.diff.read"),
|
||||
"legacy.git.force_push": CapabilityContract(
|
||||
"legacy.git.force_push",
|
||||
forbidden=("protected_ref_bypass",),
|
||||
),
|
||||
"legacy.deploy.production": CapabilityContract(
|
||||
"legacy.deploy.production",
|
||||
permissions=("production_deploy",),
|
||||
dependencies=("owner_approval", "deployment_controller"),
|
||||
health_requirements=("approval_envelope_verified", "runtime_acceptance_available"),
|
||||
),
|
||||
"legacy.secret.transform": CapabilityContract("legacy.secret.transform"),
|
||||
"legacy.secure_file_transformer": CapabilityContract("legacy.secure_file_transformer"),
|
||||
"legacy.runtime.shadow": CapabilityContract("legacy.runtime.shadow"),
|
||||
"legacy.review.security": CapabilityContract("legacy.review.security"),
|
||||
"legacy.review.api-contract": CapabilityContract("legacy.review.api-contract"),
|
||||
"legacy.gpu.train": CapabilityContract("legacy.gpu.train"),
|
||||
"agent.runtime.hermes": CapabilityContract(
|
||||
"agent.runtime.hermes",
|
||||
dependencies=("hermes_runtime",),
|
||||
health_requirements=("worker_liveness",),
|
||||
),
|
||||
"agent.runtime.openclaw": CapabilityContract(
|
||||
"agent.runtime.openclaw",
|
||||
dependencies=("openclaw_runtime",),
|
||||
health_requirements=("worker_liveness",),
|
||||
),
|
||||
"engineering.autonomous": CapabilityContract(
|
||||
"engineering.autonomous",
|
||||
dependencies=("agent_task_control_plane",),
|
||||
health_requirements=("worker_liveness",),
|
||||
),
|
||||
"release:candidate": CapabilityContract(
|
||||
"release:candidate",
|
||||
min_version=1,
|
||||
max_version=1,
|
||||
permissions=("prepare_release",),
|
||||
forbidden=(
|
||||
"production_deploy",
|
||||
"submit_owner_approval",
|
||||
"tag_publish",
|
||||
"baseline_registry_finalize",
|
||||
"runtime_switch",
|
||||
),
|
||||
dependencies=(
|
||||
"credential_projection",
|
||||
"trusted_builder",
|
||||
),
|
||||
health_requirements=(
|
||||
"authenticated_probe",
|
||||
"release_api_health",
|
||||
"worker_liveness",
|
||||
),
|
||||
),
|
||||
}
|
||||
|
||||
|
||||
def capability_contract(name: object) -> CapabilityContract | None:
|
||||
"""Return the contract for one canonical or registered capability name."""
|
||||
|
||||
text = _clean(name)
|
||||
direct = _CAPABILITY_CONTRACTS.get(text)
|
||||
if direct is not None:
|
||||
return direct
|
||||
mapped = _TASK_ALIASES.get(text) or _WORKER_ALIASES.get(text)
|
||||
if mapped is None or len(mapped) != 1:
|
||||
return None
|
||||
return _CAPABILITY_CONTRACTS.get(next(iter(mapped)))
|
||||
|
||||
|
||||
def registered_capability(name: object, *, worker: bool = False) -> bool:
|
||||
"""Fail-closed validation used by registration/task admission."""
|
||||
|
||||
text = _clean(name)
|
||||
if "*" in text:
|
||||
return False
|
||||
aliases = _WORKER_ALIASES if worker else _TASK_ALIASES
|
||||
mapped = aliases.get(text)
|
||||
if mapped is None:
|
||||
return text in _CAPABILITY_CONTRACTS
|
||||
return all(item in _CAPABILITY_CONTRACTS for item in mapped)
|
||||
|
||||
|
||||
def validate_capability_contract_registry() -> dict[str, object]:
|
||||
"""Prove alias/contract closure and contract safety invariants."""
|
||||
|
||||
alias_targets: set[str] = set()
|
||||
for mapping in (_TASK_ALIASES, _WORKER_ALIASES):
|
||||
for values in mapping.values():
|
||||
alias_targets.update(values)
|
||||
missing = sorted(alias_targets - set(_CAPABILITY_CONTRACTS))
|
||||
invalid_versions = sorted(
|
||||
name
|
||||
for name, contract in _CAPABILITY_CONTRACTS.items()
|
||||
if contract.min_version < 1 or contract.max_version < contract.min_version
|
||||
)
|
||||
permission_forbidden_overlap = sorted(
|
||||
name
|
||||
for name, contract in _CAPABILITY_CONTRACTS.items()
|
||||
if set(contract.permissions) & set(contract.forbidden)
|
||||
)
|
||||
return {
|
||||
"schema": "aios.capability-contract-registry.v1",
|
||||
"contract_count": len(_CAPABILITY_CONTRACTS),
|
||||
"missing_alias_targets": missing,
|
||||
"invalid_versions": invalid_versions,
|
||||
"permission_forbidden_overlap": permission_forbidden_overlap,
|
||||
"valid": not missing and not invalid_versions and not permission_forbidden_overlap,
|
||||
}
|
||||
|
||||
|
||||
def capability_contract_snapshot() -> dict[str, dict[str, Any]]:
|
||||
return {
|
||||
name: contract.as_dict()
|
||||
for name, contract in sorted(_CAPABILITY_CONTRACTS.items())
|
||||
}
|
||||
|
||||
|
||||
_UNKNOWN_PREFIX = "exact:"
|
||||
|
||||
|
||||
@@ -98,7 +272,11 @@ def normalize_required_capabilities(values: Iterable[object]) -> frozenset[str]:
|
||||
if "*" in text:
|
||||
result.add(_UNKNOWN_PREFIX + text)
|
||||
else:
|
||||
result.update(_TASK_ALIASES.get(text, frozenset({_UNKNOWN_PREFIX + text})))
|
||||
mapped = _TASK_ALIASES.get(text)
|
||||
if mapped is None or any(item not in _CAPABILITY_CONTRACTS for item in mapped):
|
||||
result.add(_UNKNOWN_PREFIX + text)
|
||||
else:
|
||||
result.update(mapped)
|
||||
return frozenset(result)
|
||||
|
||||
|
||||
@@ -111,7 +289,7 @@ def normalize_worker_capabilities(values: Iterable[object]) -> frozenset[str]:
|
||||
if "*" in text:
|
||||
continue
|
||||
mapped = _WORKER_ALIASES.get(text)
|
||||
if mapped is not None:
|
||||
if mapped is not None and all(item in _CAPABILITY_CONTRACTS for item in mapped):
|
||||
result.update(mapped)
|
||||
if {
|
||||
"governed.git.write.branch",
|
||||
|
||||
@@ -1,9 +1,10 @@
|
||||
"""Loopback-only governed Bridge v2 HTTP gateway with durable handlers."""
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass, replace
|
||||
from dataclasses import asdict, dataclass, replace
|
||||
from datetime import datetime, timedelta, timezone
|
||||
import base64
|
||||
import hmac
|
||||
import json
|
||||
import os
|
||||
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
|
||||
@@ -12,6 +13,8 @@ import re
|
||||
import sqlite3
|
||||
import threading
|
||||
from typing import Any, Mapping
|
||||
import urllib.error
|
||||
import urllib.request
|
||||
|
||||
from aios_authority_git_executor import AuthorityGitExecutor
|
||||
from aios_chatgpt_bridge_v2 import BridgeV2Adapter, BridgeV2Error, OPERATION_REQUIRED_FIELDS
|
||||
@@ -19,7 +22,9 @@ from aios_continuation_executor import AuthorityGitAdapter, ContinuationCommit,
|
||||
from aios_continuation_lease import WriterLeaseRequestV2
|
||||
from aios_control_plane_snapshot import build_control_plane_snapshot
|
||||
from aios_gitops_job_controller import GitOpsJobController, JobError, JobState, JobStore
|
||||
from aios_governed_task_bootstrap import BootstrapError, GovernedTaskBootstrap
|
||||
from aios_governed_task_bootstrap import BootstrapError, GovernedTaskBinding, GovernedTaskBootstrap, _legacy_safe, _task_key
|
||||
from aios_online_agent_execution import ExecutionProjectionError, OnlineTaskProjectionClient
|
||||
from aios_online_agent_plan_store import ExecutionPlanError, ExecutionPlanStore
|
||||
from governed_production_release import build_executor_config
|
||||
|
||||
MAX_BODY = 2 * 1024 * 1024
|
||||
@@ -27,6 +32,8 @@ ROUTES = {
|
||||
"/v2/runtime": "getRuntimeStatus",
|
||||
"/v2/capabilities": "getGovernedWriteCapabilities",
|
||||
"/v2/governed-task": "createGovernedTask",
|
||||
"/v2/execution-envelope": "projectOnlineAgentExecution",
|
||||
"/v2/execution-plan": "registerOnlineAgentExecutionPlan",
|
||||
"/v2/writer-lease": "acquireWriterLease",
|
||||
"/v2/writer-lease/bootstrap": "bootstrapWriterLease",
|
||||
"/v2/writer-lease/renew": "renewWriterLease",
|
||||
@@ -62,6 +69,9 @@ class RuntimeConfig:
|
||||
registry_root: Path = Path("/var/lib/aios-gitops/runtime/registry-projection")
|
||||
baseline_path: Path = Path("/var/lib/aios-gitops-runtime-anchor/baseline/current.json")
|
||||
health_url: str = "http://127.0.0.1:18927/health"
|
||||
agent_task_api_url: str = "http://127.0.0.1:18940"
|
||||
agent_task_control_token_path: Path = Path("/run/credentials/aios-chatgpt-bridge-v2-staging.service/agent_task_control_token")
|
||||
online_task_projection_enabled: bool = False
|
||||
|
||||
|
||||
def _safe(value: str) -> str:
|
||||
@@ -74,6 +84,16 @@ class GovernedRuntime:
|
||||
config.state_root.mkdir(parents=True, exist_ok=True)
|
||||
self.authority = AuthorityGitExecutor(config.authoritative_git_dir, config.state_root / "worktrees")
|
||||
self.bootstrap = GovernedTaskBootstrap(self.authority, config.state_root, remote_name=config.remote_name, token_path=config.token_path, ssh_key_path=config.ssh_key_path, api_base=config.api_base)
|
||||
self.online_task_projector = (
|
||||
OnlineTaskProjectionClient(config.agent_task_api_url, config.agent_task_control_token_path)
|
||||
if config.online_task_projection_enabled
|
||||
else None
|
||||
)
|
||||
self.projection_outbox_root = config.state_root / "online-task-projection-outbox"
|
||||
self.projection_outbox_root.mkdir(parents=True, exist_ok=True)
|
||||
self.execution_plan_store = ExecutionPlanStore(config.state_root / "online-agent-execution-plans")
|
||||
self._projection_locks_guard = threading.Lock()
|
||||
self._projection_locks: dict[str, threading.RLock] = {}
|
||||
self.runtimes: dict[tuple[str, int], tuple[Any, GitOpsJobController, ContinuationExecutor]] = {}
|
||||
self.controllers: dict[str, GitOpsJobController] = {}
|
||||
self._recovery_lock = threading.RLock()
|
||||
@@ -93,6 +113,54 @@ class GovernedRuntime:
|
||||
askpass.write_text('#!/bin/sh\ncase "$1" in *Username*) printf "%s\\n" "$GITEA_MACHINE_ACTOR" ;; *) cat "$CREDENTIAL_FILE" ;; esac\n', encoding="utf-8")
|
||||
askpass.chmod(0o700)
|
||||
os.environ.setdefault("GIT_ASKPASS", str(askpass)); os.environ.setdefault("GIT_ASKPASS_REQUIRE", "force"); os.environ.setdefault("GIT_TERMINAL_PROMPT", "0"); os.environ.setdefault("GITEA_MACHINE_ACTOR", "aios-autonomy-agent")
|
||||
if self.online_task_projector is not None:
|
||||
threading.Thread(target=self._projection_reconciler, name="online-task-projection-reconciler", daemon=True).start()
|
||||
|
||||
def _projection_lock(self, task_id: str) -> threading.RLock:
|
||||
guard = getattr(self, "_projection_locks_guard", None)
|
||||
if guard is None:
|
||||
guard = threading.Lock()
|
||||
self._projection_locks_guard = guard
|
||||
self._projection_locks = {}
|
||||
with guard:
|
||||
return self._projection_locks.setdefault(task_id, threading.RLock())
|
||||
|
||||
def _projection_outbox_path(self, task_id: str) -> Path:
|
||||
return self.projection_outbox_root / (_task_key(task_id) + ".json")
|
||||
|
||||
def _write_projection_outbox(self, task_id: str, projection: Mapping[str, Any], binding: Any) -> Path:
|
||||
path = self._projection_outbox_path(task_id)
|
||||
temporary = path.with_name(path.name + f".{os.getpid()}.{threading.get_ident()}.tmp")
|
||||
temporary.write_text(json.dumps({"task": dict(projection), "binding": asdict(binding)}, sort_keys=True, separators=(",", ":")), encoding="utf-8")
|
||||
temporary.replace(path)
|
||||
return path
|
||||
|
||||
def _deliver_projection_outbox(self, path: Path) -> dict[str, Any]:
|
||||
value = json.loads(path.read_text(encoding="utf-8"))
|
||||
if not isinstance(value, Mapping) or not isinstance(value.get("task"), Mapping) or not isinstance(value.get("binding"), Mapping):
|
||||
raise RuntimeError("ONLINE_TASK_PROJECTION_OUTBOX_INVALID")
|
||||
binding = self.bootstrap.execution_binding(GovernedTaskBinding(**dict(value["binding"])))
|
||||
if self.online_task_projector is None:
|
||||
raise RuntimeError("ONLINE_TASK_PROJECTION_CONTROL_UNAVAILABLE")
|
||||
created = self.online_task_projector.create(dict(value["task"]), binding=binding)
|
||||
path.unlink(missing_ok=True)
|
||||
return created
|
||||
|
||||
def _projection_reconciler(self) -> None:
|
||||
while True:
|
||||
for path in sorted(self.projection_outbox_root.glob("*.json")):
|
||||
try:
|
||||
value = json.loads(path.read_text(encoding="utf-8"))
|
||||
binding = value.get("binding") if isinstance(value, Mapping) else None
|
||||
task_id = str(binding.get("task_id") or "") if isinstance(binding, Mapping) else ""
|
||||
if not task_id:
|
||||
continue
|
||||
with self._projection_lock(task_id):
|
||||
if path.exists():
|
||||
self._deliver_projection_outbox(path)
|
||||
except Exception:
|
||||
continue
|
||||
threading.Event().wait(5)
|
||||
|
||||
def runtime_status(self) -> dict[str, Any]:
|
||||
return {
|
||||
@@ -113,7 +181,10 @@ class GovernedRuntime:
|
||||
}
|
||||
|
||||
def _job_root(self, task_id: str) -> Path:
|
||||
return self.config.state_root / "jobs" / _safe(task_id)
|
||||
parent = self.config.state_root / "jobs"
|
||||
keyed = parent / _task_key(task_id)
|
||||
legacy = parent / _legacy_safe(task_id)
|
||||
return keyed if keyed.exists() or not legacy.exists() else legacy
|
||||
|
||||
def _build_runtime(
|
||||
self,
|
||||
@@ -443,7 +514,85 @@ class GovernedRuntime:
|
||||
return result
|
||||
|
||||
def create_task(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
result = self.bootstrap.create(request); self.bindings[result["task_id"]] = result; return result
|
||||
online_request = request.get("online_task")
|
||||
if online_request is not None and not isinstance(online_request, Mapping):
|
||||
raise ValueError("ONLINE_TASK_REQUEST_INVALID")
|
||||
execution_request = request.get("execution")
|
||||
if execution_request is not None and not isinstance(execution_request, Mapping):
|
||||
raise ValueError("EXECUTION_REQUEST_INVALID")
|
||||
online_task_projector = getattr(self, "online_task_projector", None)
|
||||
projection_required = (
|
||||
online_request is not None
|
||||
or execution_request is not None
|
||||
or request.get("require_online_task_projection") is True
|
||||
)
|
||||
if online_task_projector is None:
|
||||
if projection_required:
|
||||
raise RuntimeError("ONLINE_TASK_PROJECTION_CONTROL_UNAVAILABLE")
|
||||
return self.bootstrap.create(request)
|
||||
online = dict(online_request or {})
|
||||
online.setdefault("goal", str(request.get("description") or "Governed engineering task"))
|
||||
online.setdefault("instructions", "Execute only the exact governed binding and return server-verifiable evidence.")
|
||||
online.setdefault("success_criteria", [{"id": "governed", "description": "Complete the exact governed task lifecycle."}])
|
||||
online.setdefault("required_capabilities", ["git.read", "git.test", "governed.write.branch", "governed.write.file"])
|
||||
online.setdefault("priority", int(request.get("priority") or 0))
|
||||
online.setdefault("max_attempts", int(request.get("max_attempts") or 10))
|
||||
normalized = dict(request)
|
||||
normalized["online_task"] = online
|
||||
normalized["execution"] = dict(execution_request or {})
|
||||
normalized["require_online_task_projection"] = True
|
||||
result = self.bootstrap.create(normalized)
|
||||
self.bindings[result["task_id"]] = result
|
||||
projection = self.bootstrap.online_task_projection(
|
||||
result["task_id"],
|
||||
online,
|
||||
normalized["execution"],
|
||||
expected_job_id=str(result.get("job_id") or "") or None,
|
||||
)
|
||||
durable_binding = self.bootstrap.get_binding(result["task_id"])
|
||||
with self._projection_lock(result["task_id"]):
|
||||
path = self._write_projection_outbox(result["task_id"], projection, durable_binding)
|
||||
created = self._deliver_projection_outbox(path)
|
||||
return {
|
||||
**result,
|
||||
"online_task_projection": {
|
||||
"status": "PROJECTED",
|
||||
"task_id": created["task_id"],
|
||||
"job_id": result["job_id"],
|
||||
"task_status": created["status"],
|
||||
"execution_digest": projection["metadata"]["execution"]["execution_digest"],
|
||||
},
|
||||
}
|
||||
|
||||
def project_online_agent_execution(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
capabilities = request.get("required_capabilities") or []
|
||||
if not isinstance(capabilities, list) or not all(isinstance(item, str) for item in capabilities):
|
||||
raise ValueError("CAPABILITY_SCOPE_INVALID")
|
||||
execution = request.get("execution")
|
||||
if execution is not None and not isinstance(execution, Mapping):
|
||||
raise ValueError("EXECUTION_REQUEST_INVALID")
|
||||
task_id = str(request["task_id"])
|
||||
durable = self.bootstrap.get_binding(task_id)
|
||||
binding = self.bootstrap.execution_binding(durable)
|
||||
expected_job_id = str(request.get("job_id") or "") or None
|
||||
if expected_job_id and expected_job_id != binding.job_id:
|
||||
raise BootstrapError("JOB_BINDING_CONFLICT")
|
||||
if execution is None:
|
||||
planned = self.execution_plan_store.resolve(binding)
|
||||
if planned is not None:
|
||||
execution = planned["execution"]
|
||||
return self.bootstrap.execution_envelope(
|
||||
task_id,
|
||||
execution,
|
||||
required_capabilities=tuple(capabilities),
|
||||
expected_job_id=expected_job_id,
|
||||
)
|
||||
|
||||
def register_online_agent_execution_plan(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
task_id = str(request["task_id"])
|
||||
durable = self.bootstrap.get_binding(task_id)
|
||||
binding = self.bootstrap.execution_binding(durable)
|
||||
return self.execution_plan_store.register(binding, request)
|
||||
|
||||
def acquire_lease(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
binding, controller, _ = self._runtime(request)
|
||||
@@ -1340,7 +1489,19 @@ class GovernedRuntime:
|
||||
controller = self._controller_for_job(job_id)
|
||||
row = controller.store.read(job_id)
|
||||
if action == "get":
|
||||
return controller.store._receipt(row, duplicate=True).as_dict()
|
||||
response = controller.store._receipt(row, duplicate=True).as_dict()
|
||||
try:
|
||||
with controller.store._connect() as connection:
|
||||
connection.row_factory = sqlite3.Row
|
||||
leases = connection.execute(
|
||||
"SELECT lease_id,job_id,branch,lineage_anchor_oid,current_head_oid,generation,"
|
||||
"lease_expires_at,state,revision FROM branch_leases WHERE job_id=? ORDER BY revision DESC",
|
||||
(job_id,),
|
||||
).fetchall()
|
||||
except sqlite3.Error as exc:
|
||||
raise RuntimeError("JOB_DB_UNAVAILABLE") from exc
|
||||
response["branch_leases"] = [dict(item) for item in leases]
|
||||
return response
|
||||
if action == "resume":
|
||||
return controller.resume(job_id).as_dict()
|
||||
return self._executor_from_row(row, controller).reconcile(job_id)
|
||||
@@ -1405,6 +1566,8 @@ class GovernedRuntime:
|
||||
def handlers(self) -> dict[str, Any]:
|
||||
return {
|
||||
"createGovernedTask": self.create_task,
|
||||
"projectOnlineAgentExecution": self.project_online_agent_execution,
|
||||
"registerOnlineAgentExecutionPlan": self.register_online_agent_execution_plan,
|
||||
"acquireWriterLease": self.acquire_lease,
|
||||
"bootstrapWriterLease": self.bootstrap_writer_lease,
|
||||
"renewWriterLease": self.renew_writer_lease,
|
||||
@@ -1434,6 +1597,8 @@ class _HTTPError(RuntimeError): pass
|
||||
|
||||
class BridgeHandler(BaseHTTPRequestHandler):
|
||||
adapter: BridgeV2Adapter | None = None
|
||||
control_token: str | None = None
|
||||
agent_task_api_url: str = "http://127.0.0.1:18940"
|
||||
def log_message(self, *_args: Any) -> None: return
|
||||
def _json(self, value: Any, status: int = 200) -> None:
|
||||
raw = json.dumps(value, sort_keys=True, separators=(",", ":")).encode(); self.send_response(status); self.send_header("Content-Type", "application/json"); self.send_header("Content-Length", str(len(raw))); self.end_headers(); self.wfile.write(raw)
|
||||
@@ -1451,7 +1616,50 @@ class BridgeHandler(BaseHTTPRequestHandler):
|
||||
allowed = set(OPERATION_REQUIRED_FIELDS[operation]) | {"schema", "job_id", "lease_seconds", "create_target_branch", "worktree_mode", "cas_ref_update_required", "expected_branch_oid", "generation", "branch_mode"}
|
||||
if operation == "commitFilesGoverned": allowed.add("files")
|
||||
allowed.update({"owner_approval", "execute"})
|
||||
if operation == "projectOnlineAgentExecution":
|
||||
allowed.update({"execution", "required_capabilities"})
|
||||
if operation == "registerOnlineAgentExecutionPlan":
|
||||
allowed.update({"execution", "source_head_oid", "expected_previous_plan_digest"})
|
||||
if operation == "createGovernedTask":
|
||||
allowed.update({"online_task", "execution", "require_online_task_projection", "priority", "max_attempts", "parent_task_id", "parent_job_id"})
|
||||
if set(body) - allowed: raise _HTTPError("UNKNOWN_FIELDS")
|
||||
def _authorized(self, operation: str, body: Mapping[str, Any]) -> bool:
|
||||
header = self.headers.get("Authorization", "")
|
||||
if not header.startswith("Bearer "):
|
||||
return False
|
||||
supplied = header[7:]
|
||||
if self.control_token and hmac.compare_digest(supplied, self.control_token):
|
||||
return True
|
||||
poll = urllib.request.Request(
|
||||
self.agent_task_api_url.rstrip("/") + "/v1/tasks/poll",
|
||||
method="POST",
|
||||
data=b'{"limit":1}',
|
||||
headers={"Authorization": header, "Content-Type": "application/json", "Accept": "application/json"},
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(poll, timeout=5) as response:
|
||||
principal = json.loads(response.read().decode("utf-8"))
|
||||
except (OSError, urllib.error.URLError, urllib.error.HTTPError, ValueError):
|
||||
return False
|
||||
agent_id = str(principal.get("agent_id") or "") if isinstance(principal, Mapping) else ""
|
||||
if not agent_id:
|
||||
return False
|
||||
claim_bound = {"acquireWriterLease", "bootstrapWriterLease", "renewWriterLease", "commitFileGoverned", "commitFilesGoverned", "projectOnlineAgentExecution", "getGitOpsJob"}
|
||||
if operation not in claim_bound:
|
||||
return False
|
||||
task_id = str(body.get("task_id") or "")
|
||||
if not task_id:
|
||||
return False
|
||||
task_request = urllib.request.Request(
|
||||
self.agent_task_api_url.rstrip("/") + "/v1/tasks/" + task_id,
|
||||
headers={"Authorization": header, "Accept": "application/json"},
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(task_request, timeout=5) as response:
|
||||
task = json.loads(response.read().decode("utf-8"))
|
||||
except (OSError, urllib.error.URLError, urllib.error.HTTPError, ValueError):
|
||||
return False
|
||||
return isinstance(task, Mapping) and task.get("task_id") == task_id and task.get("claimed_by") == agent_id and task.get("status") in {"CLAIMED", "RUNNING"}
|
||||
def do_GET(self) -> None:
|
||||
if self.adapter is None: self._json({"error": "RUNTIME_UNAVAILABLE"}, 503); return
|
||||
if self.path == "/v2/runtime": self._json(self.adapter.getRuntimeStatus()); return
|
||||
@@ -1462,14 +1670,23 @@ class BridgeHandler(BaseHTTPRequestHandler):
|
||||
operation = ROUTES.get(self.path)
|
||||
if operation is None or operation in ("getRuntimeStatus", "getGovernedWriteCapabilities") or self.adapter is None: self._json({"error": "NOT_FOUND"}, 404); return
|
||||
try:
|
||||
body = self._body(); self._allowed(operation, body); self._json(getattr(self.adapter, operation)(body))
|
||||
body = self._body()
|
||||
except _HTTPError as exc: self._json({"error": str(exc)}, 413 if str(exc) == "REQUEST_TOO_LARGE" else 400); return
|
||||
if not self._authorized(operation, body): self._json({"error": "BRIDGE_CALLER_UNAUTHORIZED"}, 401); return
|
||||
try:
|
||||
self._allowed(operation, body); self._json(getattr(self.adapter, operation)(body))
|
||||
except _HTTPError as exc: self._json({"error": str(exc)}, 413 if str(exc) == "REQUEST_TOO_LARGE" else 400)
|
||||
except (BootstrapError, ValueError, BridgeV2Error) as exc: self._json({"error": str(exc)}, 409 if "CONFLICT" in str(exc) else 400)
|
||||
except (BootstrapError, ExecutionProjectionError, ExecutionPlanError, ValueError, BridgeV2Error) as exc: self._json({"error": str(exc)}, 409 if any(token in str(exc) for token in ("CONFLICT", "DRIFT", "CAS")) else 400)
|
||||
except Exception as exc: self._json({"error": str(exc) or type(exc).__name__}, 409 if any(token in str(exc) for token in ("CONFLICT", "EXPECTATION", "MONOTONIC")) else 503)
|
||||
|
||||
|
||||
def serve(adapter: BridgeV2Adapter, host: str = "127.0.0.1", port: int = 18892) -> None:
|
||||
BridgeHandler.adapter = adapter; ThreadingHTTPServer((host, port), BridgeHandler).serve_forever()
|
||||
def serve(adapter: BridgeV2Adapter, host: str = "127.0.0.1", port: int = 18892, *, control_token_path: Path, agent_task_api_url: str) -> None:
|
||||
if not control_token_path.is_file() or control_token_path.is_symlink():
|
||||
raise RuntimeError("BRIDGE_CONTROL_CREDENTIAL_UNAVAILABLE")
|
||||
control_token = control_token_path.read_text(encoding="utf-8").strip()
|
||||
if len(control_token) < 32:
|
||||
raise RuntimeError("BRIDGE_CONTROL_CREDENTIAL_INVALID")
|
||||
BridgeHandler.adapter = adapter; BridgeHandler.control_token = control_token; BridgeHandler.agent_task_api_url = agent_task_api_url; ThreadingHTTPServer((host, port), BridgeHandler).serve_forever()
|
||||
|
||||
|
||||
def main() -> None:
|
||||
@@ -1488,11 +1705,21 @@ def main() -> None:
|
||||
remote_name=os.environ.get("AIOS_GITEA_REMOTE", "gitea-http"),
|
||||
api_base=os.environ.get("AIOS_GITEA_API_BASE", "https://g.cnnetmkt.top/git/api/v1"),
|
||||
target_version=os.environ.get("AIOS_CONTROL_PLANE_TARGET_VERSION", "V2R2C1"),
|
||||
agent_task_api_url=os.environ.get("AIOS_AGENT_TASK_API_URL", "http://127.0.0.1:18940"),
|
||||
agent_task_control_token_path=Path(
|
||||
os.environ.get(
|
||||
"AIOS_AGENT_TASK_CONTROL_TOKEN_FILE",
|
||||
str(Path(os.environ.get("CREDENTIALS_DIRECTORY", "/run/credentials/aios-chatgpt-bridge-v2-staging.service")) / "agent_task_control_token"),
|
||||
)
|
||||
),
|
||||
online_task_projection_enabled=os.environ.get("AIOS_ONLINE_TASK_PROJECTION_ENABLED", "1") == "1",
|
||||
)
|
||||
runtime = GovernedRuntime(config)
|
||||
serve(
|
||||
BridgeV2Adapter(runtime_status=runtime.runtime_status(), handlers=runtime.handlers()),
|
||||
port=int(os.environ.get("AIOS_BRIDGE_V2_PORT", "18892")),
|
||||
control_token_path=config.agent_task_control_token_path,
|
||||
agent_task_api_url=config.agent_task_api_url,
|
||||
)
|
||||
|
||||
|
||||
|
||||
@@ -157,7 +157,7 @@ class DetachedGitWorkspace:
|
||||
def materialize_files(self, files: Mapping[str, bytes]) -> None:
|
||||
dirty = self._dirty_paths()
|
||||
expected_paths = set(files)
|
||||
if dirty and dirty != expected_paths:
|
||||
if dirty and not dirty.issubset(expected_paths):
|
||||
raise ContinuationExecutionError("WORKTREE_DIRTY_SCOPE_MISMATCH")
|
||||
for relative, content in files.items():
|
||||
target = (self.path / relative).resolve()
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@@ -2,13 +2,20 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass, asdict
|
||||
from contextlib import contextmanager
|
||||
import fcntl
|
||||
import hashlib
|
||||
import json
|
||||
from pathlib import Path
|
||||
import re
|
||||
from typing import Any, Mapping
|
||||
|
||||
from aios_gitops_job_controller import GitOpsJobController, JobStore
|
||||
from aios_gitops_job_controller import GitOpsJobController, JobStore, sha256_json
|
||||
from aios_online_agent_execution import (
|
||||
GovernedExecutionBinding,
|
||||
project_execution_envelope,
|
||||
project_online_task,
|
||||
)
|
||||
|
||||
OID_RE = re.compile(r"^[0-9a-f]{40}$")
|
||||
BRANCH_RE = re.compile(r"^agent/[A-Za-z0-9][A-Za-z0-9._/-]{2,179}$")
|
||||
@@ -36,35 +43,161 @@ class GovernedTaskBinding:
|
||||
return {"task_id": self.task_id, "repository": self.repository, "base_oid": self.base_oid, "target_branch": self.target_branch, "expected_branch_oid": self.expected_branch_oid, "generation": self.generation, "branch_mode": self.branch_mode, "create_target_branch": self.branch_mode == "CREATE_NEW", "worktree_mode": "detached_at_base_oid" if self.branch_mode == "CREATE_NEW" else "detached_at_expected_head", "cas_ref_update_required": True, "idempotent": idempotent, "state": "BOUND", "job_id": self.job_id}
|
||||
|
||||
|
||||
def _safe(value: str) -> str:
|
||||
def _legacy_safe(value: str) -> str:
|
||||
return re.sub(r"[^A-Za-z0-9_.-]+", "-", value).strip("-")[:120] or "task"
|
||||
|
||||
|
||||
def _task_key(value: str) -> str:
|
||||
readable = _legacy_safe(value)[:72]
|
||||
return f"{readable}-{hashlib.sha256(value.encode('utf-8')).hexdigest()[:24]}"
|
||||
|
||||
|
||||
class GovernedTaskBootstrap:
|
||||
def __init__(self, authority: Any, root: Path, *, remote_name: str, token_path: Path, ssh_key_path: Path, api_base: str = "https://g.cnnetmkt.top/git/api/v1") -> None:
|
||||
self.authority = authority; self.root = Path(root); self.remote_name = remote_name; self.token_path = Path(token_path); self.ssh_key_path = Path(ssh_key_path); self.api_base = api_base; self.task_meta_root = self.root / "task-bindings"; self.job_root = self.root / "jobs"; self.task_meta_root.mkdir(parents=True, exist_ok=True); self.job_root.mkdir(parents=True, exist_ok=True)
|
||||
self.authority = authority; self.root = Path(root); self.remote_name = remote_name; self.token_path = Path(token_path); self.ssh_key_path = Path(ssh_key_path); self.api_base = api_base; self.task_meta_root = self.root / "task-bindings"; self.job_root = self.root / "jobs"; self.lock_root = self.root / "task-locks"; self.task_meta_root.mkdir(parents=True, exist_ok=True); self.job_root.mkdir(parents=True, exist_ok=True); self.lock_root.mkdir(parents=True, exist_ok=True)
|
||||
|
||||
def _store(self, task_id: str) -> tuple[JobStore, GitOpsJobController]:
|
||||
root = self.job_root / _safe(task_id); store = JobStore(root, payload_root=root / "payloads"); return store, GitOpsJobController(store)
|
||||
legacy = self.job_root / _legacy_safe(task_id)
|
||||
keyed = self.job_root / _task_key(task_id)
|
||||
root = legacy if legacy.is_dir() and not keyed.exists() else keyed
|
||||
store = JobStore(root, payload_root=root / "payloads"); return store, GitOpsJobController(store)
|
||||
|
||||
def _metadata(self, task_id: str) -> Path:
|
||||
return self.task_meta_root / (_safe(task_id) + ".json")
|
||||
return self.task_meta_root / (_task_key(task_id) + ".json")
|
||||
|
||||
def _legacy_metadata(self, task_id: str) -> Path:
|
||||
return self.task_meta_root / (_legacy_safe(task_id) + ".json")
|
||||
|
||||
@contextmanager
|
||||
def _task_lock(self, task_id: str):
|
||||
path = self.lock_root / (_task_key(task_id) + ".lock")
|
||||
with path.open("a+b") as handle:
|
||||
fcntl.flock(handle.fileno(), fcntl.LOCK_EX)
|
||||
try:
|
||||
yield
|
||||
finally:
|
||||
fcntl.flock(handle.fileno(), fcntl.LOCK_UN)
|
||||
|
||||
@contextmanager
|
||||
def _branch_lock(self, repository: str, target_branch: str):
|
||||
key = hashlib.sha256((repository + "\0" + target_branch).encode("utf-8")).hexdigest()
|
||||
path = self.lock_root / ("branch-" + key + ".lock")
|
||||
with path.open("a+b") as handle:
|
||||
fcntl.flock(handle.fileno(), fcntl.LOCK_EX)
|
||||
try:
|
||||
yield
|
||||
finally:
|
||||
fcntl.flock(handle.fileno(), fcntl.LOCK_UN)
|
||||
|
||||
def _read_existing(self, task_id: str) -> GovernedTaskBinding | None:
|
||||
path = self._metadata(task_id)
|
||||
if not path.exists():
|
||||
path = self._legacy_metadata(task_id)
|
||||
if not path.is_file() or path.is_symlink(): return None
|
||||
return GovernedTaskBinding(**json.loads(path.read_text(encoding="utf-8")))
|
||||
binding = GovernedTaskBinding(**json.loads(path.read_text(encoding="utf-8")))
|
||||
if binding.task_id != task_id:
|
||||
raise BootstrapError("TASK_BINDING_KEY_COLLISION")
|
||||
return binding
|
||||
|
||||
def get_binding(self, task_id: str) -> GovernedTaskBinding:
|
||||
"""Return the exact durable binding; never synthesize missing lineage."""
|
||||
existing = self._read_existing(str(task_id))
|
||||
if existing is None or not existing.job_id:
|
||||
raise BootstrapError("TASK_EXECUTION_PROJECTION_MISSING")
|
||||
return existing
|
||||
|
||||
def execution_binding(self, binding: GovernedTaskBinding) -> GovernedExecutionBinding:
|
||||
expected = binding.expected_branch_oid
|
||||
mode = binding.branch_mode
|
||||
if mode == "CREATE_NEW":
|
||||
current = self.authority.read_ref("refs/heads/" + binding.target_branch)
|
||||
if current is not None:
|
||||
store, _ = self._store(binding.task_id)
|
||||
with store._connect() as connection:
|
||||
rows = connection.execute(
|
||||
"SELECT current_head_oid,generation,state,lineage_anchor_oid "
|
||||
"FROM branch_leases WHERE job_id=? AND branch=? AND state='ACTIVE' "
|
||||
"ORDER BY revision DESC",
|
||||
(binding.job_id, binding.target_branch),
|
||||
).fetchall()
|
||||
if len(rows) != 1:
|
||||
raise BootstrapError("SAME_JOB_CONTINUATION_LEASE_INVALID")
|
||||
row = rows[0]
|
||||
if (
|
||||
str(row["current_head_oid"] or "") != current
|
||||
or int(row["generation"] or 0) != binding.generation
|
||||
or str(row["lineage_anchor_oid"] or "") != binding.base_oid
|
||||
):
|
||||
raise BootstrapError("SAME_JOB_CONTINUATION_BINDING_INVALID")
|
||||
expected = current
|
||||
mode = "CONTINUE_EXISTING"
|
||||
return GovernedExecutionBinding.from_mapping(
|
||||
{
|
||||
"task_id": binding.task_id,
|
||||
"job_id": binding.job_id,
|
||||
"repository": binding.repository,
|
||||
"base_oid": binding.base_oid,
|
||||
"expected_branch_oid": expected,
|
||||
"target_branch": binding.target_branch,
|
||||
"branch_mode": mode,
|
||||
"generation": binding.generation,
|
||||
"idempotency_key": binding.idempotency_key,
|
||||
}
|
||||
)
|
||||
|
||||
def execution_envelope(
|
||||
self,
|
||||
task_id: str,
|
||||
execution_request: Mapping[str, Any] | None = None,
|
||||
*,
|
||||
required_capabilities: tuple[str, ...] = (),
|
||||
expected_job_id: str | None = None,
|
||||
) -> dict[str, Any]:
|
||||
binding = self.get_binding(task_id)
|
||||
if expected_job_id and binding.job_id != expected_job_id:
|
||||
raise BootstrapError("JOB_BINDING_CONFLICT")
|
||||
return project_execution_envelope(
|
||||
self.execution_binding(binding),
|
||||
execution_request,
|
||||
required_capabilities=required_capabilities,
|
||||
)
|
||||
|
||||
def online_task_projection(
|
||||
self,
|
||||
task_id: str,
|
||||
task_request: Mapping[str, Any],
|
||||
execution_request: Mapping[str, Any] | None = None,
|
||||
*,
|
||||
expected_job_id: str | None = None,
|
||||
) -> dict[str, Any]:
|
||||
binding = self.get_binding(task_id)
|
||||
if expected_job_id and binding.job_id != expected_job_id:
|
||||
raise BootstrapError("JOB_BINDING_CONFLICT")
|
||||
return project_online_task(self.execution_binding(binding), task_request, execution_request)
|
||||
|
||||
def _write(self, binding: GovernedTaskBinding) -> None:
|
||||
path = self._metadata(binding.task_id); temporary = path.with_suffix(".tmp"); temporary.write_text(json.dumps(asdict(binding), sort_keys=True), encoding="utf-8"); temporary.replace(path)
|
||||
|
||||
def create(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
task_id = str(request.get("task_id") or "")
|
||||
if not task_id:
|
||||
raise BootstrapError("TASK_BINDING_INVALID")
|
||||
repository = str(request.get("repository") or "")
|
||||
target_branch = str(request.get("target_branch") or "")
|
||||
with self._task_lock(task_id):
|
||||
with self._branch_lock(repository, target_branch):
|
||||
return self._create_locked(request)
|
||||
|
||||
def _create_locked(self, request: Mapping[str, Any]) -> dict[str, Any]:
|
||||
values = {key: str(request[key]) for key in ("task_id", "repository", "base_oid", "target_branch", "risk", "description", "idempotency_key")}
|
||||
if values["repository"] != "aios/aios" or not OID_RE.fullmatch(values["base_oid"]): raise BootstrapError("TASK_BINDING_INVALID")
|
||||
if not BRANCH_RE.fullmatch(values["target_branch"]): raise BootstrapError("TARGET_BRANCH_INVALID")
|
||||
existing = self._read_existing(values["task_id"])
|
||||
if existing is not None:
|
||||
if any(getattr(existing, key) != values[key] for key in ("repository", "base_oid", "target_branch", "risk", "description")): raise BootstrapError("TASK_BINDING_CONFLICT")
|
||||
if any(getattr(existing, key) != values[key] for key in ("repository", "base_oid", "target_branch", "risk", "description", "idempotency_key")): raise BootstrapError("TASK_BINDING_CONFLICT")
|
||||
optional = {"expected_branch_oid": existing.expected_branch_oid, "branch_mode": existing.branch_mode, "generation": existing.generation}
|
||||
for key, durable in optional.items():
|
||||
if key in request and str(request[key]) != str(durable): raise BootstrapError("TASK_BINDING_CONFLICT")
|
||||
return existing.response(idempotent=True)
|
||||
if not self.authority.commit_exists(values["base_oid"]): raise BootstrapError("BASE_COMMIT_NOT_FOUND")
|
||||
current = self.authority.read_ref("refs/heads/" + values["target_branch"])
|
||||
@@ -74,8 +207,61 @@ class GovernedTaskBootstrap:
|
||||
if expected != values["base_oid"]: raise BootstrapError("EXPECTED_BASE_CONFLICT")
|
||||
else:
|
||||
if expected != current: raise BootstrapError("EXISTING_BRANCH_OID_CONFLICT")
|
||||
mode, generation = "CONTINUE_EXISTING", int(request.get("generation", 2))
|
||||
if generation < 1: raise BootstrapError("GENERATION_INVALID")
|
||||
parent_task_id = str(request.get("parent_task_id") or "")
|
||||
parent_job_id = str(request.get("parent_job_id") or "")
|
||||
if not parent_task_id or not parent_job_id:
|
||||
raise BootstrapError("EXISTING_BRANCH_OWNERSHIP_REQUIRED")
|
||||
parent = self._read_existing(parent_task_id)
|
||||
if parent is None or parent.job_id != parent_job_id or parent.repository != values["repository"] or parent.target_branch != values["target_branch"]:
|
||||
raise BootstrapError("EXISTING_BRANCH_OWNERSHIP_CONFLICT")
|
||||
store, _ = self._store(parent_task_id)
|
||||
with store._connect() as connection:
|
||||
rows = connection.execute(
|
||||
"SELECT current_head_oid,generation,state FROM branch_leases WHERE job_id=? AND branch=? AND state='ACTIVE' ORDER BY revision DESC",
|
||||
(parent_job_id, values["target_branch"]),
|
||||
).fetchall()
|
||||
if len(rows) != 1 or str(rows[0]["current_head_oid"] or "") != current:
|
||||
raise BootstrapError("EXISTING_BRANCH_OWNERSHIP_CONFLICT")
|
||||
mode, generation = "CONTINUE_EXISTING", int(rows[0]["generation"]) + 1
|
||||
if "generation" in request and int(request["generation"]) != generation:
|
||||
raise BootstrapError("GENERATION_BINDING_CONFLICT")
|
||||
online_request = request.get("online_task")
|
||||
execution_request = request.get("execution")
|
||||
projection_required = (
|
||||
online_request is not None
|
||||
or execution_request is not None
|
||||
or request.get("require_online_task_projection") is True
|
||||
)
|
||||
if projection_required:
|
||||
if not isinstance(online_request, Mapping):
|
||||
raise BootstrapError("ONLINE_TASK_REQUEST_INVALID")
|
||||
if execution_request is not None and not isinstance(execution_request, Mapping):
|
||||
raise BootstrapError("EXECUTION_REQUEST_INVALID")
|
||||
provisional_job_id = "job-" + sha256_json(
|
||||
{"client_idempotency_key": values["idempotency_key"]}
|
||||
)[:32]
|
||||
try:
|
||||
provisional_binding = GovernedExecutionBinding.from_mapping(
|
||||
{
|
||||
"task_id": values["task_id"],
|
||||
"job_id": provisional_job_id,
|
||||
"repository": values["repository"],
|
||||
"base_oid": values["base_oid"],
|
||||
"expected_branch_oid": expected,
|
||||
"target_branch": values["target_branch"],
|
||||
"branch_mode": mode,
|
||||
"generation": generation,
|
||||
"idempotency_key": values["idempotency_key"],
|
||||
}
|
||||
)
|
||||
project_online_task(
|
||||
provisional_binding,
|
||||
dict(online_request),
|
||||
dict(execution_request or {}),
|
||||
)
|
||||
except Exception as exc:
|
||||
raise BootstrapError("EXECUTION_BINDING_INVALID") from exc
|
||||
|
||||
payload = json.dumps(values, sort_keys=True, separators=(",", ":")).encode()
|
||||
_, controller = self._store(values["task_id"])
|
||||
receipt = controller.submit({"idempotency_key": values["idempotency_key"], "campaign_id": values["task_id"], "task_id": values["task_id"], "repository": values["repository"], "operation": "GOVERNED_TASK_CREATE", "branch": values["target_branch"], "expected_branch_oid": expected, "payload_digest": hashlib.sha256(payload).hexdigest(), "lineage_anchor_oid": values["base_oid"]}, payload=payload)
|
||||
|
||||
@@ -3,6 +3,15 @@ from __future__ import annotations
|
||||
import hashlib, hmac, json, os, re, subprocess, sys, time, urllib.parse, urllib.request, uuid
|
||||
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
|
||||
from pathlib import Path
|
||||
|
||||
from aios_caller_admission_projection import (
|
||||
CALLER,
|
||||
SCOPE,
|
||||
GOVERNED_TASK_SCOPE,
|
||||
PROJECTED_SCOPES,
|
||||
CallerProjectionClient,
|
||||
CallerProjectionError,
|
||||
)
|
||||
from baseline_authority import BaselineAuthority, BaselineAuthorityError
|
||||
from typing import Any
|
||||
|
||||
@@ -89,6 +98,11 @@ SILENT_RELEASE_PUBLIC_KEYS = Path(os.getenv(
|
||||
'AIOS_SILENT_RELEASE_PUBLIC_KEYS',
|
||||
'/etc/aios-gitops/silent-release-owner-public-keys.json',
|
||||
))
|
||||
CALLER_PROJECTION_SOCKET = Path(os.getenv(
|
||||
'AIOS_OWNER_AUTH_BROKER_SOCKET',
|
||||
'/run/aios-owner-authorization-broker/broker.sock',
|
||||
))
|
||||
CALLER_PROJECTION_CLIENT = CallerProjectionClient(CALLER_PROJECTION_SOCKET)
|
||||
monitor = ActionMonitor()
|
||||
|
||||
BASE_MODELS = [
|
||||
@@ -112,6 +126,9 @@ READ_ONLY_GET_PATHS = frozenset({
|
||||
'/admin/ai-evolution/api/data',
|
||||
'/api/release-consistency', '/api/tag-publication-status',
|
||||
'/api/release-acceptance',
|
||||
'/api/governed-write-capabilities', '/api/control-plane-snapshot',
|
||||
'/api/caller-capabilities',
|
||||
'/api/online-agent-task', '/api/online-agent-task-events',
|
||||
'/getReleaseCandidate', '/getReleasePlan', '/getReleaseAcceptance',
|
||||
'/getApprovalEnvelope', '/getReleaseSaga', '/getCapabilityStatus',
|
||||
'/getRecoveryStatus',
|
||||
@@ -124,6 +141,25 @@ SILENT_RELEASE_POST_PATHS = frozenset({
|
||||
'/startDeployment',
|
||||
})
|
||||
|
||||
GOVERNED_EXECUTION_CAPABILITIES = frozenset({
|
||||
"governed.write.branch",
|
||||
"governed.write.file",
|
||||
})
|
||||
|
||||
|
||||
def requires_governed_execution(value: Mapping[str, Any]) -> bool:
|
||||
capabilities = value.get("required_capabilities")
|
||||
if capabilities is None:
|
||||
return False
|
||||
if not isinstance(capabilities, list):
|
||||
return True
|
||||
return bool(
|
||||
GOVERNED_EXECUTION_CAPABILITIES.intersection(
|
||||
{str(item) for item in capabilities}
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def reject_non_read_http(handler, method: str, path: str) -> bool:
|
||||
if method == 'GET':
|
||||
return False
|
||||
@@ -407,14 +443,113 @@ class H(BaseHTTPRequestHandler):
|
||||
except Exception: return {'_raw':raw}
|
||||
def actor(self):
|
||||
auth=self.headers.get('Authorization',''); tok=auth[7:].strip() if auth.startswith('Bearer ') else self.headers.get('X-AIOS-Token','')
|
||||
rec=authenticate(tok)
|
||||
if rec: return {'ok':True,'account':safe(rec.get('account') or rec.get('name') or 'token-user'),'role':rec.get('role','user'),'scopes':set(rec.get('scopes') or [])}
|
||||
return {'ok':False,'account':'anonymous','role':'anonymous','scopes':set()}
|
||||
def need(self, scope):
|
||||
a=self.actor()
|
||||
if not a['ok']: self.sendj({'error':'unauthorized','hint':'Bearer token required'},401); return None
|
||||
if '*' not in a['scopes'] and scope not in a['scopes']: self.sendj({'error':'forbidden','required_scope':scope,'account':a['account']},403); return None
|
||||
try:
|
||||
rec=authenticate(tok)
|
||||
except Exception:
|
||||
rec=None
|
||||
if not isinstance(rec, dict) or not rec:
|
||||
return {
|
||||
'ok': False,
|
||||
'account': 'anonymous',
|
||||
'role': 'anonymous',
|
||||
'scopes': set(),
|
||||
'static_scopes': set(),
|
||||
'projection': None,
|
||||
'credential_loaded': False,
|
||||
}
|
||||
scopes = set(rec.get('scopes') or [])
|
||||
return {
|
||||
'ok': True,
|
||||
'account': safe(rec.get('account') or rec.get('name') or 'token-user'),
|
||||
'role': rec.get('role', 'user'),
|
||||
'scopes': scopes,
|
||||
'static_scopes': set(scopes),
|
||||
'projection': None,
|
||||
'credential_loaded': False,
|
||||
}
|
||||
def need(self, scope, *, projection_request=None):
|
||||
a = self.actor()
|
||||
if not a['ok']:
|
||||
self.sendj({'error': 'unauthorized', 'hint': 'Bearer token required'}, 401)
|
||||
return None
|
||||
if a['account'] == CALLER and scope in PROJECTED_SCOPES:
|
||||
try:
|
||||
result = CALLER_PROJECTION_CLIENT.rehydrate(
|
||||
a['account'],
|
||||
scope,
|
||||
projection_request,
|
||||
)
|
||||
except CallerProjectionError as exc:
|
||||
self.sendj(
|
||||
{
|
||||
'error': 'caller_projection_denied',
|
||||
'projection_state': 'DENIED',
|
||||
'detail': exc.code,
|
||||
'account': a['account'],
|
||||
},
|
||||
exc.status,
|
||||
)
|
||||
return None
|
||||
projection = result.public
|
||||
a['scopes'].update(projection.get('effective_scopes') or [])
|
||||
a['projection'] = projection
|
||||
a['credential_loaded'] = result.credential_loaded
|
||||
if '*' not in a['scopes'] and scope not in a['scopes']:
|
||||
self.sendj(
|
||||
{
|
||||
'error': 'forbidden',
|
||||
'required_scope': scope,
|
||||
'account': a['account'],
|
||||
},
|
||||
403,
|
||||
)
|
||||
return None
|
||||
return a
|
||||
def caller_capabilities(self):
|
||||
actor = self.actor()
|
||||
if not actor['ok']:
|
||||
return self.sendj(
|
||||
{'error': 'unauthorized', 'hint': 'Bearer token required'},
|
||||
401,
|
||||
)
|
||||
effective = set(actor['static_scopes'])
|
||||
projection_state = 'NOT_REQUIRED'
|
||||
credential_state = 'NOT_REQUIRED'
|
||||
rehydration_state = 'NOT_APPLICABLE'
|
||||
admitted = True
|
||||
projection_error = None
|
||||
if actor['account'] == CALLER:
|
||||
try:
|
||||
result = CALLER_PROJECTION_CLIENT.rehydrate(
|
||||
actor['account'],
|
||||
GOVERNED_TASK_SCOPE,
|
||||
{},
|
||||
)
|
||||
except CallerProjectionError as exc:
|
||||
admitted = False
|
||||
projection_state = 'DENIED'
|
||||
credential_state = 'NOT_REHYDRATED'
|
||||
rehydration_state = 'FAILED'
|
||||
projection_error = exc.code
|
||||
else:
|
||||
effective.update(result.public.get('effective_scopes') or [])
|
||||
projection_state = result.public['projection_state']
|
||||
credential_state = result.public['credential_state']
|
||||
rehydration_state = result.public['rehydration_state']
|
||||
return self.sendj(
|
||||
{
|
||||
'schema': 'aios.caller-capability-introspection.v1',
|
||||
'caller_identity': actor['account'],
|
||||
'admitted': admitted,
|
||||
'effective_scopes': sorted(effective),
|
||||
'projection_state': projection_state,
|
||||
'credential_state': credential_state,
|
||||
'rehydration_state': rehydration_state,
|
||||
'projection_error': projection_error,
|
||||
'secret_exposed': False,
|
||||
}
|
||||
)
|
||||
|
||||
def bridge_release(self, path, data):
|
||||
route = {
|
||||
'/createTagApprovalRequest': '/v2/tag/approval-request',
|
||||
@@ -446,24 +581,183 @@ class H(BaseHTTPRequestHandler):
|
||||
return self.sendj(body, exc.code)
|
||||
except (urllib.error.URLError, TimeoutError) as exc:
|
||||
return self.sendj({'error': 'GOVERNED_BRIDGE_UNAVAILABLE', 'detail': type(exc).__name__}, 503)
|
||||
def do_POST(self):
|
||||
path = urllib.parse.urlparse(self.path).path
|
||||
if path in {
|
||||
'/createTagApprovalRequest', '/requestOrCreateTag', '/getTagPublicationStatus',
|
||||
'/buildProductionArtifact', '/executeApprovedProductionRelease',
|
||||
'/getReleaseConsistency', '/runReleaseAcceptance',
|
||||
}:
|
||||
actor = self.need('tag:create')
|
||||
if not actor:
|
||||
return
|
||||
return self.bridge_release(path, self.body())
|
||||
if reject_non_read_http(self, 'POST', path):
|
||||
return
|
||||
return self.sendj({'error': 'not_found', 'path': path}, 404)
|
||||
def bridge_v2_json(self, route, data=None, method='POST'):
|
||||
base = os.getenv('AIOS_BRIDGE_V2_URL', 'http://127.0.0.1:18892').rstrip('/')
|
||||
payload = None if data is None else json.dumps(
|
||||
data, sort_keys=True, separators=(',', ':')
|
||||
).encode('utf-8')
|
||||
request = urllib.request.Request(
|
||||
base + route,
|
||||
data=payload,
|
||||
headers={'Content-Type': 'application/json'} if payload is not None else {},
|
||||
method=method,
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(request, timeout=30) as response:
|
||||
result = json.loads(response.read().decode('utf-8'))
|
||||
return self.sendj(result, response.status)
|
||||
except urllib.error.HTTPError as exc:
|
||||
try:
|
||||
body = json.loads(exc.read().decode('utf-8'))
|
||||
except Exception:
|
||||
body = {'error': 'bridge_http_error'}
|
||||
return self.sendj(body, exc.code)
|
||||
except (urllib.error.URLError, TimeoutError) as exc:
|
||||
return self.sendj(
|
||||
{'error': 'GOVERNED_BRIDGE_UNAVAILABLE', 'detail': type(exc).__name__},
|
||||
503,
|
||||
)
|
||||
|
||||
def control_plane_snapshot(self):
|
||||
"""Build a non-sensitive snapshot locally when Bridge v2 has no snapshot HTTP route."""
|
||||
base = os.getenv('AIOS_BRIDGE_V2_URL', 'http://127.0.0.1:18892').rstrip('/')
|
||||
|
||||
def bridge_get(route):
|
||||
request = urllib.request.Request(
|
||||
base + route,
|
||||
headers={'Accept': 'application/json'},
|
||||
method='GET',
|
||||
)
|
||||
with urllib.request.urlopen(request, timeout=15) as response:
|
||||
if response.status != 200:
|
||||
raise RuntimeError('bridge_read_http_' + str(response.status))
|
||||
value = json.loads(response.read().decode('utf-8'))
|
||||
if not isinstance(value, dict):
|
||||
raise RuntimeError('bridge_read_not_object')
|
||||
return value
|
||||
|
||||
try:
|
||||
runtime = bridge_get('/v2/runtime')
|
||||
capabilities = bridge_get('/v2/capabilities')
|
||||
repository = repository_snapshot(
|
||||
git_dir=AUTHORITATIVE_GIT_DIR,
|
||||
repository_identity=REPOSITORY_IDENTITY,
|
||||
)
|
||||
if not repository.get('repository_available'):
|
||||
return self.sendj({
|
||||
'error': 'AUTHORITATIVE_REPOSITORY_UNAVAILABLE',
|
||||
'query_error': repository.get('query_error'),
|
||||
}, 503)
|
||||
baseline = dict(BASELINE_AUTHORITY.read())
|
||||
except BaselineAuthorityError as exc:
|
||||
return self.sendj({
|
||||
'error': 'RECONCILIATION_REQUIRED',
|
||||
'detail': str(exc),
|
||||
}, 503)
|
||||
except urllib.error.HTTPError as exc:
|
||||
return self.sendj({
|
||||
'error': 'GOVERNED_BRIDGE_UNAVAILABLE',
|
||||
'detail': 'HTTP_' + str(exc.code),
|
||||
}, 503)
|
||||
except (urllib.error.URLError, TimeoutError, OSError, ValueError, RuntimeError, json.JSONDecodeError) as exc:
|
||||
return self.sendj({
|
||||
'error': 'GOVERNED_BRIDGE_UNAVAILABLE',
|
||||
'detail': type(exc).__name__,
|
||||
}, 503)
|
||||
|
||||
legacy = capabilities.get('legacy_writes_enabled')
|
||||
if legacy is None:
|
||||
legacy = capabilities.get('legacy_global_writes_enabled')
|
||||
payload = {
|
||||
'schema': 'aios.control-plane-snapshot.aggregate.v1',
|
||||
'snapshot_source': 'model-api-local-aggregate',
|
||||
'version': {
|
||||
'target_release': runtime.get('control_plane_release_version'),
|
||||
},
|
||||
'runtime': {
|
||||
'bridge_schema_version': runtime.get('bridge_schema_version'),
|
||||
'control_plane_release_version': runtime.get('control_plane_release_version'),
|
||||
'runtime_release_commit': runtime.get('runtime_release_commit'),
|
||||
'runtime_path': runtime.get('runtime_path'),
|
||||
'artifact_sha256': runtime.get('artifact_sha256'),
|
||||
'drift_status': runtime.get('drift_status'),
|
||||
'deployment_state': runtime.get('deployment_state'),
|
||||
},
|
||||
'repository': {
|
||||
'identity': runtime.get('repository_identity') or runtime.get('repository') or REPOSITORY_IDENTITY,
|
||||
'authority_git_dir': runtime.get('authoritative_git_dir') or repository.get('authoritative_git_dir'),
|
||||
'authority_main_oid': repository.get('authoritative_head_oid'),
|
||||
},
|
||||
'production': {
|
||||
'baseline_version': baseline.get('current_baseline'),
|
||||
'target_oid': baseline.get('target_oid'),
|
||||
'deployment_state': baseline.get('status'),
|
||||
'baseline_generation': baseline.get('generation'),
|
||||
'baseline_sha256': baseline.get('authority_sha256'),
|
||||
},
|
||||
'task': {},
|
||||
'capabilities': {
|
||||
'legacy_writes_enabled': legacy,
|
||||
'governed_writes_enabled': capabilities.get('governed_writes_enabled'),
|
||||
'force_push': capabilities.get('force_push'),
|
||||
'main_write': capabilities.get('main_write'),
|
||||
'tag_write': capabilities.get('tag_write'),
|
||||
'baseline_write': capabilities.get('baseline_write'),
|
||||
'production_write': capabilities.get('production_write'),
|
||||
},
|
||||
'credential_broker': {
|
||||
'status': 'SERVER_SIDE_ONLY',
|
||||
'provider': 'BridgeV2RuntimeCredentialProjection',
|
||||
},
|
||||
'secret_material_exposed': False,
|
||||
}
|
||||
return self.sendj(payload)
|
||||
|
||||
def agent_task_control(self, method, route, data=None):
|
||||
token_path = Path(os.getenv(
|
||||
'AIOS_AGENT_TASK_CONTROL_TOKEN_FILE',
|
||||
'/run/credentials/aios-gitops-model-api-v2r2c1-staging.service/agent_task_control_token',
|
||||
))
|
||||
if token_path.is_symlink() or not token_path.is_file():
|
||||
return self.sendj({'error': 'AGENT_TASK_CONTROL_CREDENTIAL_UNAVAILABLE'}, 503)
|
||||
token = token_path.read_text(encoding='utf-8').strip()
|
||||
if len(token) < 32:
|
||||
return self.sendj({'error': 'AGENT_TASK_CONTROL_CREDENTIAL_INVALID'}, 503)
|
||||
base = os.getenv('AIOS_AGENT_TASK_API_URL', 'http://127.0.0.1:18940').rstrip('/')
|
||||
payload = None if data is None else json.dumps(
|
||||
data, sort_keys=True, separators=(',', ':')
|
||||
).encode('utf-8')
|
||||
headers = {'Authorization': 'Bearer ' + token, 'Accept': 'application/json'}
|
||||
if payload is not None:
|
||||
headers['Content-Type'] = 'application/json'
|
||||
request = urllib.request.Request(
|
||||
base + route, data=payload, headers=headers, method=method
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(request, timeout=30) as response:
|
||||
result = json.loads(response.read().decode('utf-8'))
|
||||
return self.sendj(result, response.status)
|
||||
except urllib.error.HTTPError as exc:
|
||||
try:
|
||||
body = json.loads(exc.read().decode('utf-8'))
|
||||
except Exception:
|
||||
body = {'error': 'agent_task_http_error'}
|
||||
return self.sendj(body, exc.code)
|
||||
except (urllib.error.URLError, TimeoutError) as exc:
|
||||
return self.sendj(
|
||||
{'error': 'AGENT_TASK_API_UNAVAILABLE', 'detail': type(exc).__name__},
|
||||
503,
|
||||
)
|
||||
|
||||
def do_GET(self):
|
||||
u=urllib.parse.urlparse(self.path); path=u.path.rstrip('/') if u.path != '/' else u.path
|
||||
if path not in READ_ONLY_GET_PATHS: return self.sendj({'error':'not_found','path':path},404)
|
||||
q=urllib.parse.parse_qs(u.query)
|
||||
if path == '/api/governed-write-capabilities':
|
||||
return self.bridge_v2_json('/v2/capabilities', method='GET')
|
||||
if path == '/api/control-plane-snapshot':
|
||||
return self.control_plane_snapshot()
|
||||
if path == '/api/caller-capabilities':
|
||||
return self.caller_capabilities()
|
||||
if path in ('/api/online-agent-task', '/api/online-agent-task-events'):
|
||||
actor = self.need('agent-task:control')
|
||||
if not actor:
|
||||
return
|
||||
task_id = (q.get('task_id') or [''])[0]
|
||||
if not re.fullmatch(r'[A-Za-z0-9][A-Za-z0-9._:-]{2,179}', task_id):
|
||||
return self.sendj({'error': 'task_id_invalid'}, 400)
|
||||
suffix = '/events' if path.endswith('-events') else ''
|
||||
return self.agent_task_control('GET', '/v1/tasks/' + task_id + suffix)
|
||||
try:
|
||||
if path in ('/health','/api/health'):
|
||||
return self.sendj({
|
||||
@@ -700,12 +994,68 @@ class H(BaseHTTPRequestHandler):
|
||||
return self.sendj({'error':'not_found','path':path},404)
|
||||
def do_POST(self):
|
||||
path=urllib.parse.urlparse(self.path).path.rstrip('/')
|
||||
governed_routes = {
|
||||
'/createGovernedTask': ('/v2/governed-task', 'branch:create'),
|
||||
'/acquireWriterLease': ('/v2/writer-lease', 'file:write'),
|
||||
'/commitFileGoverned': ('/v2/commit-file', 'file:write'),
|
||||
'/commitFilesGoverned': ('/v2/commit-files', 'file:write'),
|
||||
'/getGitOpsJob': ('/v2/job', 'file:write'),
|
||||
'/resumeGitOpsJob': ('/v2/job/resume', 'file:write'),
|
||||
'/reconcileGitOpsJob': ('/v2/job/reconcile', 'file:write'),
|
||||
'/createTagApprovalRequest': ('/v2/tag/approval-request', 'tag:create'),
|
||||
'/requestOrCreateTag': ('/v2/tag/publish', 'tag:create'),
|
||||
'/getTagPublicationStatus': ('/v2/tag/status', 'tag:create'),
|
||||
'/buildProductionArtifact': ('/v2/release/artifact', 'tag:create'),
|
||||
'/executeApprovedProductionRelease': ('/v2/release/execute', 'tag:create'),
|
||||
'/getReleaseConsistency': ('/v2/release/consistency', 'tag:create'),
|
||||
'/runReleaseAcceptance': ('/v2/release/acceptance', 'tag:create'),
|
||||
}
|
||||
if path in governed_routes:
|
||||
route, scope = governed_routes[path]
|
||||
if path == '/createGovernedTask' and self.actor()['account'] == CALLER:
|
||||
scope = GOVERNED_TASK_SCOPE
|
||||
actor = self.need(scope)
|
||||
if not actor:
|
||||
return
|
||||
return self.bridge_v2_json(route, self.body(), method='POST')
|
||||
if path in {
|
||||
'/createOnlineAgentTask', '/registerOnlineAgent',
|
||||
'/verifyOnlineAgentTask', '/cancelOnlineAgentTask',
|
||||
}:
|
||||
actor = self.need('agent-task:control')
|
||||
if not actor:
|
||||
return
|
||||
data = self.body()
|
||||
if path == '/createOnlineAgentTask':
|
||||
if requires_governed_execution(data):
|
||||
return self.sendj(
|
||||
{
|
||||
'error': 'governed_execution_binding_required',
|
||||
'hint': 'use createGovernedTask',
|
||||
},
|
||||
409,
|
||||
)
|
||||
return self.agent_task_control('POST', '/v1/tasks', data)
|
||||
if path == '/registerOnlineAgent':
|
||||
return self.agent_task_control('POST', '/v1/agents/register', data)
|
||||
task_id = str(data.pop('task_id', '') or '')
|
||||
if not re.fullmatch(r'[A-Za-z0-9][A-Za-z0-9._:-]{2,179}', task_id):
|
||||
return self.sendj({'error': 'task_id_invalid'}, 400)
|
||||
suffix = '/verify' if path == '/verifyOnlineAgentTask' else '/cancel'
|
||||
return self.agent_task_control('POST', '/v1/tasks/' + task_id + suffix, data)
|
||||
if reject_non_read_http(self, 'POST', path): return
|
||||
data=self.body()
|
||||
try:
|
||||
if path == '/createReleaseCandidate':
|
||||
a=self.need('release:candidate')
|
||||
if not a: return
|
||||
a = self.need(
|
||||
'release:candidate',
|
||||
projection_request={
|
||||
'source_commit': str(data.get('source_commit') or ''),
|
||||
'version': str(data.get('version') or ''),
|
||||
},
|
||||
)
|
||||
if not a:
|
||||
return
|
||||
return self.sendj(
|
||||
silent_release_service().create_release_candidate(data)
|
||||
)
|
||||
@@ -996,7 +1346,7 @@ def api_spec():
|
||||
return {
|
||||
'version': 'V2R2C1-IMMUTABLE-STAGING',
|
||||
'openapi': '/api/openapi.yaml',
|
||||
'read': ['GET /api/base','GET /api/ver','GET /api/diff','GET /api/patch','GET /api/model','GET /api/run','GET /getReleaseCandidate','GET /getApprovalEnvelope','GET /getReleaseSaga','GET /admin/ai-evolution'],
|
||||
'read': ['GET /api/caller-capabilities','GET /api/base','GET /api/ver','GET /api/diff','GET /api/patch','GET /api/model','GET /api/run','GET /getReleaseCandidate','GET /getApprovalEnvelope','GET /getReleaseSaga','GET /admin/ai-evolution'],
|
||||
'write': ['POST /api/branch','POST /api/file','POST /api/commit','POST /api/patch','POST /api/tag','POST /api/rollback','POST /api/model','PUT /api/model-priority','POST /api/nl','POST /createReleaseCandidate','POST /submitOwnerApproval','POST /startDeployment'],
|
||||
'policy': {
|
||||
'bearer_required_for_write': True,
|
||||
@@ -1011,6 +1361,9 @@ def api_spec():
|
||||
'silent_release_shadow_default': True,
|
||||
'runtime_mode': 'IMMUTABLE_STAGING',
|
||||
'governed_release_routing': 'bridge-v2',
|
||||
'caller_admission_projection': 'durable-owner-policy-server-side-ack',
|
||||
'caller_projection_scope': SCOPE,
|
||||
'caller_projection_caller': CALLER,
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
@@ -179,10 +179,7 @@ def _agent_runtime_input(value: object) -> dict[str, Any]:
|
||||
if not isinstance(value, Mapping):
|
||||
_error("AGENT_RUNTIME_INPUT_REQUIRED")
|
||||
raw = dict(value)
|
||||
allowed = {
|
||||
"goal", "instructions", "success_criteria", "constraints",
|
||||
"scope", "trace_id",
|
||||
}
|
||||
allowed = {"goal", "instructions", "success_criteria", "constraints"}
|
||||
if set(raw) - allowed:
|
||||
_error("AGENT_RUNTIME_INPUT_FIELDS_INVALID")
|
||||
goal = _string(raw.get("goal"), "AGENT_RUNTIME_GOAL_REQUIRED")
|
||||
@@ -211,34 +208,11 @@ def _agent_runtime_input(value: object) -> dict[str, Any]:
|
||||
lowered = str(key).lower()
|
||||
if any(marker in lowered for marker in sensitive_markers):
|
||||
_error("AGENT_RUNTIME_SECRET_FIELD_DENY")
|
||||
|
||||
scope = raw.get("scope", {})
|
||||
if not isinstance(scope, Mapping):
|
||||
_error("AGENT_RUNTIME_SCOPE_INVALID")
|
||||
scope = {
|
||||
"scope_type": str(scope.get("scope_type") or "legacy"),
|
||||
"tenant_id": scope.get("tenant_id"),
|
||||
"project_id": scope.get("project_id"),
|
||||
}
|
||||
if scope["scope_type"] not in {"legacy", "system", "tenant"}:
|
||||
_error("AGENT_RUNTIME_SCOPE_INVALID")
|
||||
if scope["scope_type"] == "tenant":
|
||||
if not isinstance(scope["tenant_id"], str) or not scope["tenant_id"].strip():
|
||||
_error("AGENT_RUNTIME_SCOPE_INVALID")
|
||||
elif scope["tenant_id"] not in (None, "") or scope["project_id"] not in (None, ""):
|
||||
_error("AGENT_RUNTIME_SCOPE_INVALID")
|
||||
|
||||
trace_id = str(raw.get("trace_id") or "").strip()
|
||||
if not trace_id or len(trace_id) > 255:
|
||||
_error("AGENT_RUNTIME_TRACE_ID_INVALID")
|
||||
|
||||
normalized = {
|
||||
"goal": goal[:12000],
|
||||
"instructions": instructions[:32000],
|
||||
"success_criteria": normalized_criteria,
|
||||
"constraints": dict(constraints),
|
||||
"scope": scope,
|
||||
"trace_id": trace_id,
|
||||
}
|
||||
if len(_canonical(normalized).encode("utf-8")) > MAX_AGENT_RUNTIME_INPUT_BYTES:
|
||||
_error("AGENT_RUNTIME_INPUT_TOO_LARGE")
|
||||
@@ -451,12 +425,6 @@ def project_online_task(
|
||||
if adapter not in AGENT_RUNTIME_ADAPTERS:
|
||||
_error("AGENT_RUNTIME_ADAPTER_INVALID")
|
||||
runtime_input = execution_value.get("runtime_input")
|
||||
tenant_id = request.get("tenant_id")
|
||||
project_id = request.get("project_id")
|
||||
declared_scope = str(
|
||||
request.get("scope_type")
|
||||
or ("tenant" if isinstance(tenant_id, str) and tenant_id.strip() else "legacy")
|
||||
)
|
||||
if runtime_input is None:
|
||||
runtime_input = {
|
||||
"goal": request.get("goal"),
|
||||
@@ -464,20 +432,6 @@ def project_online_task(
|
||||
"success_criteria": request.get("success_criteria") or [],
|
||||
"constraints": request.get("constraints") or {},
|
||||
}
|
||||
elif not isinstance(runtime_input, Mapping):
|
||||
_error("AGENT_RUNTIME_INPUT_REQUIRED")
|
||||
runtime_input = dict(runtime_input)
|
||||
# Scope/trace are authoritative Task identity, never caller-selected runtime hints.
|
||||
runtime_input["scope"] = {
|
||||
"scope_type": declared_scope,
|
||||
"tenant_id": tenant_id,
|
||||
"project_id": project_id,
|
||||
}
|
||||
runtime_input["trace_id"] = str(
|
||||
request.get("trace_id")
|
||||
or request.get("run_id")
|
||||
or binding.task_id
|
||||
)
|
||||
execution_value["runtime_input"] = runtime_input
|
||||
execution = project_execution_envelope(
|
||||
binding,
|
||||
|
||||
@@ -1910,6 +1910,8 @@ class WorkerSupervisor:
|
||||
# engineering placeholder: never pass that stale payload back to the
|
||||
# resolver, and never claim it until a deterministic plan exists.
|
||||
pending_plan = source_capture_requires_deterministic_plan(persisted)
|
||||
if not pending_plan:
|
||||
return persisted, "persisted_verified"
|
||||
projection_input = (
|
||||
execution
|
||||
if persisted.action in {"source_capture", "agent_runtime"} and not pending_plan
|
||||
@@ -1958,25 +1960,7 @@ class WorkerSupervisor:
|
||||
phase="projection",
|
||||
)
|
||||
return persisted, "persisted_verified"
|
||||
status, projected = self.bridge.project_execution(task)
|
||||
if status == 404 or projected.get("error") == "TASK_EXECUTION_PROJECTION_MISSING":
|
||||
raise WorkerError("TASK_SKIPPED_PROJECTION_MISSING", phase="projection")
|
||||
if not 200 <= status < 300:
|
||||
raise WorkerError(
|
||||
"TASK_SKIPPED_POLICY_DENY",
|
||||
phase="projection",
|
||||
details={"api_status": status, "error": projected.get("error")},
|
||||
)
|
||||
envelope = ExecutionEnvelope.from_task(task, projected, WORKER_CAPABILITIES)
|
||||
audit(
|
||||
"TASK_EXECUTION_ENVELOPE_PROJECTED",
|
||||
task_id=envelope.task_id,
|
||||
job_id=envelope.job_id,
|
||||
generation=envelope.generation,
|
||||
action=envelope.action,
|
||||
execution_digest=envelope.digest,
|
||||
)
|
||||
return envelope, "recovered_projection"
|
||||
raise WorkerError("TASK_SKIPPED_PROJECTION_MISSING", phase="projection")
|
||||
|
||||
def process_claim(
|
||||
self,
|
||||
@@ -2448,23 +2432,21 @@ class WorkerSupervisor:
|
||||
write_worker_liveness("POLL_OK", visible_count=len(tasks))
|
||||
audit("TASK_POLL", visible_count=len(tasks))
|
||||
except WorkerError as exc:
|
||||
write_worker_liveness(
|
||||
"POLL_FAILED",
|
||||
error=exc.code,
|
||||
)
|
||||
write_worker_liveness("POLL_FAILED", error=exc.code)
|
||||
audit("TASK_POLL_FAILED", error=exc.code, phase=exc.phase, details=exc.details)
|
||||
return 10
|
||||
dynamic_tasks: list[tuple[dict[str, Any], ExecutionEnvelope]] = []
|
||||
dynamic_tasks: list[tuple[dict[str, Any], ExecutionEnvelope, str]] = []
|
||||
for task in tasks:
|
||||
try:
|
||||
envelope, source = self.resolve_envelope(task)
|
||||
metadata = task.get("metadata")
|
||||
execution = metadata.get("execution") if isinstance(metadata, Mapping) else None
|
||||
if not isinstance(execution, Mapping):
|
||||
raise WorkerError("TASK_SKIPPED_PROJECTION_MISSING", phase="projection")
|
||||
envelope = ExecutionEnvelope.from_task(task, worker_capabilities=WORKER_CAPABILITIES)
|
||||
source = "pending_plan" if source_capture_requires_deterministic_plan(envelope) else "persisted_verified"
|
||||
except WorkerError as exc:
|
||||
event = "TASK_SKIPPED_ENVELOPE_INVALID"
|
||||
if exc.code in {
|
||||
"TASK_SKIPPED_PROJECTION_MISSING",
|
||||
"TASK_SKIPPED_POLICY_DENY",
|
||||
"TASK_SKIPPED_PENDING_DETERMINISTIC_PLAN",
|
||||
}:
|
||||
if exc.code in {"TASK_SKIPPED_PROJECTION_MISSING", "TASK_SKIPPED_POLICY_DENY", "TASK_SKIPPED_PENDING_DETERMINISTIC_PLAN"}:
|
||||
event = exc.code
|
||||
elif exc.code == "CAPABILITY_MISMATCH_DENY":
|
||||
event = "TASK_SKIPPED_CAPABILITY_MISMATCH"
|
||||
@@ -2472,16 +2454,11 @@ class WorkerSupervisor:
|
||||
continue
|
||||
audit("TASK_VISIBLE_TO_WORKER_CLAIMABLE", task_id=task.get("task_id"), envelope_source=source, action=envelope.action)
|
||||
if envelope.adapter not in WORKER_ALLOWED_ADAPTERS:
|
||||
audit(
|
||||
"TASK_SKIPPED_ADAPTER_MISMATCH",
|
||||
task_id=task.get("task_id"),
|
||||
adapter=envelope.adapter,
|
||||
allowed_adapters=sorted(WORKER_ALLOWED_ADAPTERS),
|
||||
)
|
||||
audit("TASK_SKIPPED_ADAPTER_MISMATCH", task_id=task.get("task_id"), adapter=envelope.adapter, allowed_adapters=sorted(WORKER_ALLOWED_ADAPTERS))
|
||||
continue
|
||||
dynamic_tasks.append((task, envelope))
|
||||
dynamic_tasks.sort(key=_claim_order)
|
||||
for task, envelope in dynamic_tasks:
|
||||
dynamic_tasks.append((task, envelope, source))
|
||||
dynamic_tasks.sort(key=lambda item: _claim_order(item[0]))
|
||||
for task, envelope, source in dynamic_tasks:
|
||||
task_id = str(task.get("task_id") or "")
|
||||
status, claim = self.agent.claim(task_id)
|
||||
if status == 409:
|
||||
@@ -2490,6 +2467,19 @@ class WorkerSupervisor:
|
||||
if status != 200:
|
||||
audit("TASK_CLAIM_FAILED", task_id=task_id, api_status=status, error=claim.get("error"))
|
||||
continue
|
||||
if source == "pending_plan":
|
||||
current_status, current_task = self.agent.get(task_id)
|
||||
if current_status == 200:
|
||||
try:
|
||||
envelope, _ = self.resolve_envelope(current_task)
|
||||
except WorkerError as exc:
|
||||
audit("TASK_PLAN_RESOLUTION_FAILED", task_id=task_id, reason=exc.code, phase=exc.phase, details=exc.details)
|
||||
self.process_claim(task, claim, envelope)
|
||||
return 0
|
||||
else:
|
||||
audit("TASK_READ_AFTER_CLAIM_FAILED", task_id=task_id, api_status=current_status)
|
||||
self.process_claim(task, claim, envelope)
|
||||
return 0
|
||||
self.process_claim(task, claim, envelope)
|
||||
return 0
|
||||
return 0
|
||||
|
||||
@@ -0,0 +1,82 @@
|
||||
"""Immutable release-lineage contracts used by activation and recovery."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import asdict, dataclass
|
||||
from typing import Any, Mapping
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class ReleaseLineage:
|
||||
task_id: str
|
||||
candidate_id: str
|
||||
source_commit: str
|
||||
version: str
|
||||
required_capabilities: tuple[str, ...]
|
||||
|
||||
def as_dict(self) -> dict[str, Any]:
|
||||
return asdict(self)
|
||||
|
||||
|
||||
FIXED_V2R2C2 = ReleaseLineage(
|
||||
task_id="r42-release-candidate-v2r2c2-live-20260819",
|
||||
candidate_id="r42-release-candidate-v2r2c2-20260819",
|
||||
source_commit="71e736b05fab13efc06af7a4fe2dc022c1cf9a85",
|
||||
version="V2R2C2",
|
||||
required_capabilities=("release:candidate",),
|
||||
)
|
||||
|
||||
|
||||
class ReleaseLineageError(RuntimeError):
|
||||
pass
|
||||
|
||||
|
||||
def validate_fixed_v2r2c2_task(task: Mapping[str, Any]) -> dict[str, Any]:
|
||||
"""Validate identity/source/version without mutating Task state.
|
||||
|
||||
This contract intentionally does not bind status/attempt/revision, because
|
||||
those are lifecycle state and may change through governed retry/recovery.
|
||||
"""
|
||||
|
||||
if str(task.get("task_id") or "") != FIXED_V2R2C2.task_id:
|
||||
raise ReleaseLineageError("FIXED_TASK_ID_DRIFT")
|
||||
|
||||
capabilities = task.get("required_capabilities")
|
||||
if (
|
||||
not isinstance(capabilities, list)
|
||||
or tuple(capabilities) != FIXED_V2R2C2.required_capabilities
|
||||
):
|
||||
raise ReleaseLineageError("FIXED_TASK_CAPABILITY_DRIFT")
|
||||
|
||||
constraints = task.get("constraints")
|
||||
if not isinstance(constraints, Mapping):
|
||||
raise ReleaseLineageError("FIXED_TASK_CONSTRAINTS_MISSING")
|
||||
release = constraints.get("release_candidate")
|
||||
if not isinstance(release, Mapping):
|
||||
raise ReleaseLineageError("FIXED_RELEASE_BINDING_MISSING")
|
||||
|
||||
actual = {
|
||||
"candidate_id": str(release.get("candidate_id") or ""),
|
||||
"source_commit": str(release.get("source_commit") or ""),
|
||||
"version": str(release.get("version") or ""),
|
||||
}
|
||||
expected = {
|
||||
"candidate_id": FIXED_V2R2C2.candidate_id,
|
||||
"source_commit": FIXED_V2R2C2.source_commit,
|
||||
"version": FIXED_V2R2C2.version,
|
||||
}
|
||||
if actual != expected:
|
||||
raise ReleaseLineageError(
|
||||
"FIXED_RELEASE_LINEAGE_DRIFT:"
|
||||
+ repr({"expected": expected, "actual": actual})
|
||||
)
|
||||
|
||||
return {
|
||||
"schema": "aios.release-lineage.validation.v1",
|
||||
"status": "VERIFIED",
|
||||
"lineage": FIXED_V2R2C2.as_dict(),
|
||||
"task_status": task.get("status"),
|
||||
"attempt": task.get("attempt"),
|
||||
"max_attempts": task.get("max_attempts"),
|
||||
"revision": task.get("revision"),
|
||||
}
|
||||
@@ -0,0 +1,714 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Persistent release-preparation Online Agent worker.
|
||||
|
||||
The worker is intentionally narrower than the generic engineering worker:
|
||||
it can only consume tasks whose required capability set is exactly
|
||||
{"release:candidate"} and whose payload asks for immutable release-candidate
|
||||
creation. It never receives Owner signing, deployment, tag, baseline or Git
|
||||
write credentials.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime, timezone
|
||||
import hashlib
|
||||
import json
|
||||
import os
|
||||
from pathlib import Path
|
||||
import re
|
||||
import socket
|
||||
import threading
|
||||
import time
|
||||
import urllib.error
|
||||
import urllib.request
|
||||
from typing import Any, Mapping
|
||||
from aios_autonomous_control_plane import (
|
||||
classify_failure,
|
||||
repair_for_failure,
|
||||
runtime_admission,
|
||||
)
|
||||
|
||||
AGENT_ID = os.environ.get(
|
||||
"AIOS_RELEASE_WORKER_AGENT_ID",
|
||||
"online-release-worker-cn-network-mkt-20260819",
|
||||
).strip()
|
||||
AGENT_VERSION = os.environ.get(
|
||||
"AIOS_RELEASE_WORKER_AGENT_VERSION",
|
||||
"2026-08-19.1-release-candidate",
|
||||
).strip()
|
||||
TASK_API = os.environ.get(
|
||||
"AIOS_AGENT_TASK_API_URL",
|
||||
"http://127.0.0.1:18940",
|
||||
).rstrip("/")
|
||||
RELEASE_API = os.environ.get("AIOS_RELEASE_API_URL", "").rstrip("/")
|
||||
POLL_SECONDS = int(os.environ.get("AIOS_RELEASE_WORKER_POLL_SECONDS", "3"))
|
||||
LEASE_SECONDS = int(os.environ.get("AIOS_RELEASE_WORKER_LEASE_SECONDS", "60"))
|
||||
HEARTBEAT_SECONDS = int(
|
||||
os.environ.get("AIOS_RELEASE_WORKER_HEARTBEAT_SECONDS", "15")
|
||||
)
|
||||
STATE_ROOT = Path(
|
||||
os.environ.get(
|
||||
"AIOS_RELEASE_WORKER_STATE_ROOT",
|
||||
"/var/lib/aios-online-agent-release",
|
||||
)
|
||||
)
|
||||
EXPECTED_CAPABILITIES = frozenset({"release:candidate"})
|
||||
LIVENESS_PATH = Path(
|
||||
os.environ.get(
|
||||
"AIOS_WORKER_LIVENESS_PATH",
|
||||
str(STATE_ROOT / "worker-liveness.json"),
|
||||
)
|
||||
)
|
||||
CANDIDATE_RE = re.compile(r"^[A-Za-z0-9][A-Za-z0-9._:-]{7,255}$")
|
||||
OID_RE = re.compile(r"^[0-9a-f]{40}$")
|
||||
VERSION_RE = re.compile(r"^V[0-9]+R[0-9]+C[0-9]+$")
|
||||
TASK_RE = re.compile(r"^[A-Za-z0-9][A-Za-z0-9._:-]{2,179}$")
|
||||
SECRET_MARKERS = (
|
||||
"authorization",
|
||||
"token",
|
||||
"secret",
|
||||
"credential",
|
||||
"password",
|
||||
"private_key",
|
||||
"signature",
|
||||
)
|
||||
|
||||
|
||||
class ReleaseWorkerError(RuntimeError):
|
||||
def __init__(self, code: str, *, phase: str) -> None:
|
||||
super().__init__(code)
|
||||
self.code = code
|
||||
self.phase = phase
|
||||
|
||||
|
||||
def now() -> str:
|
||||
return datetime.now(timezone.utc).isoformat()
|
||||
|
||||
|
||||
def canonical(value: object) -> bytes:
|
||||
return json.dumps(
|
||||
value,
|
||||
ensure_ascii=False,
|
||||
sort_keys=True,
|
||||
separators=(",", ":"),
|
||||
).encode("utf-8")
|
||||
|
||||
|
||||
def sha256(value: bytes) -> str:
|
||||
return hashlib.sha256(value).hexdigest()
|
||||
|
||||
|
||||
def sanitized(value: Any) -> Any:
|
||||
if isinstance(value, Mapping):
|
||||
result: dict[str, Any] = {}
|
||||
for key, inner in value.items():
|
||||
lowered = str(key).lower()
|
||||
if any(marker in lowered for marker in SECRET_MARKERS):
|
||||
result[str(key)] = "[REDACTED]"
|
||||
else:
|
||||
result[str(key)] = sanitized(inner)
|
||||
return result
|
||||
if isinstance(value, list):
|
||||
return [sanitized(item) for item in value]
|
||||
return value
|
||||
|
||||
|
||||
def audit(event: str, **fields: Any) -> None:
|
||||
STATE_ROOT.mkdir(parents=True, exist_ok=True, mode=0o750)
|
||||
path = STATE_ROOT / "events.jsonl"
|
||||
record = {
|
||||
"schema": "aios.release-worker.event.v1",
|
||||
"event": event,
|
||||
"created_at": now(),
|
||||
"agent_id": AGENT_ID,
|
||||
**sanitized(fields),
|
||||
}
|
||||
with path.open("a", encoding="utf-8") as handle:
|
||||
handle.write(json.dumps(record, sort_keys=True) + "\n")
|
||||
os.chmod(path, 0o640)
|
||||
|
||||
|
||||
def sd_notify(message: str) -> None:
|
||||
target = os.environ.get("NOTIFY_SOCKET", "")
|
||||
if not target:
|
||||
return
|
||||
if target.startswith("@"):
|
||||
address = "\0" + target[1:]
|
||||
else:
|
||||
address = target
|
||||
sock = socket.socket(socket.AF_UNIX, socket.SOCK_DGRAM)
|
||||
try:
|
||||
sock.connect(address)
|
||||
sock.sendall(message.encode("utf-8"))
|
||||
except OSError:
|
||||
# Liveness JSON remains the diagnostic source; systemd will enforce
|
||||
# WatchdogSec if notify transport is persistently broken.
|
||||
pass
|
||||
finally:
|
||||
sock.close()
|
||||
|
||||
|
||||
def write_liveness(
|
||||
state: str,
|
||||
*,
|
||||
task_id: str | None = None,
|
||||
stage: str | None = None,
|
||||
error: str | None = None,
|
||||
visible_count: int | None = None,
|
||||
) -> None:
|
||||
STATE_ROOT.mkdir(parents=True, exist_ok=True, mode=0o750)
|
||||
value = {
|
||||
"schema": "aios.online-agent.worker-liveness.v1",
|
||||
"agent_id": AGENT_ID,
|
||||
"agent_type": "release-preparation-worker",
|
||||
"agent_version": AGENT_VERSION,
|
||||
"pid": os.getpid(),
|
||||
"state": state,
|
||||
"last_seen_epoch": time.time(),
|
||||
"last_seen_at": now(),
|
||||
"poll_seconds": POLL_SECONDS,
|
||||
"heartbeat_seconds": HEARTBEAT_SECONDS,
|
||||
"api_base": TASK_API,
|
||||
"capabilities": sorted(EXPECTED_CAPABILITIES),
|
||||
"task_id": task_id,
|
||||
"stage": stage,
|
||||
"error": error,
|
||||
"visible_count": visible_count,
|
||||
}
|
||||
LIVENESS_PATH.parent.mkdir(parents=True, exist_ok=True, mode=0o750)
|
||||
temporary = LIVENESS_PATH.with_name(
|
||||
"." + LIVENESS_PATH.name + f".{os.getpid()}.tmp"
|
||||
)
|
||||
temporary.write_text(
|
||||
json.dumps(value, ensure_ascii=False, sort_keys=True) + "\n",
|
||||
encoding="utf-8",
|
||||
)
|
||||
os.chmod(temporary, 0o640)
|
||||
os.replace(temporary, LIVENESS_PATH)
|
||||
|
||||
|
||||
def load_credential(name: str) -> str:
|
||||
directory = Path(os.environ.get("CREDENTIALS_DIRECTORY", ""))
|
||||
path = directory / name
|
||||
if path.is_symlink() or not path.is_file():
|
||||
raise ReleaseWorkerError(
|
||||
"CREDENTIAL_UNAVAILABLE:" + name,
|
||||
phase="startup",
|
||||
)
|
||||
value = path.read_text(encoding="utf-8").strip()
|
||||
if len(value) < 24 or any(ch.isspace() for ch in value):
|
||||
raise ReleaseWorkerError(
|
||||
"CREDENTIAL_INVALID:" + name,
|
||||
phase="startup",
|
||||
)
|
||||
return value
|
||||
|
||||
|
||||
def http_json(
|
||||
base: str,
|
||||
method: str,
|
||||
route: str,
|
||||
token: str,
|
||||
body: Mapping[str, Any] | None = None,
|
||||
*,
|
||||
timeout: int = 30,
|
||||
) -> tuple[int, dict[str, Any]]:
|
||||
data = None if body is None else canonical(body)
|
||||
headers = {"Accept": "application/json"}
|
||||
if token:
|
||||
headers["Authorization"] = "Bearer " + token
|
||||
if data is not None:
|
||||
headers["Content-Type"] = "application/json"
|
||||
request = urllib.request.Request(
|
||||
base + route,
|
||||
method=method,
|
||||
data=data,
|
||||
headers=headers,
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(request, timeout=timeout) as response:
|
||||
raw, status = response.read(), response.status
|
||||
except urllib.error.HTTPError as exc:
|
||||
raw, status = exc.read(), exc.code
|
||||
except (urllib.error.URLError, TimeoutError) as exc:
|
||||
raise ReleaseWorkerError(
|
||||
"HTTP_TRANSPORT_UNAVAILABLE",
|
||||
phase="network",
|
||||
) from exc
|
||||
try:
|
||||
value = json.loads(raw.decode())
|
||||
except Exception as exc:
|
||||
raise ReleaseWorkerError(
|
||||
"HTTP_RESPONSE_INVALID",
|
||||
phase="network",
|
||||
) from exc
|
||||
return status, value if isinstance(value, dict) else {}
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class ReleaseRequest:
|
||||
candidate_id: str
|
||||
source_commit: str
|
||||
version: str
|
||||
projection_id: str | None
|
||||
|
||||
@classmethod
|
||||
def from_task(cls, task: Mapping[str, Any]) -> "ReleaseRequest":
|
||||
if not TASK_RE.fullmatch(str(task.get("task_id") or "")):
|
||||
raise ReleaseWorkerError("TASK_ID_INVALID", phase="validation")
|
||||
caps = task.get("required_capabilities")
|
||||
if (
|
||||
not isinstance(caps, list)
|
||||
or frozenset(caps) != EXPECTED_CAPABILITIES
|
||||
or len(caps) != 1
|
||||
):
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_CAPABILITY_SCOPE_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
metadata = task.get("metadata")
|
||||
if not isinstance(metadata, Mapping):
|
||||
metadata = {}
|
||||
constraints = task.get("constraints")
|
||||
if not isinstance(constraints, Mapping):
|
||||
constraints = {}
|
||||
release = constraints.get("release_candidate")
|
||||
if not isinstance(release, Mapping):
|
||||
release = metadata.get("release_candidate")
|
||||
if not isinstance(release, Mapping):
|
||||
release = task.get("release_candidate")
|
||||
if not isinstance(release, Mapping):
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_CANDIDATE_PAYLOAD_MISSING",
|
||||
phase="validation",
|
||||
)
|
||||
allowed = {
|
||||
"candidate_id",
|
||||
"source_commit",
|
||||
"version",
|
||||
"projection_id",
|
||||
}
|
||||
if set(release) - allowed:
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_CANDIDATE_PAYLOAD_FIELDS_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
candidate_id = str(release.get("candidate_id") or "")
|
||||
source_commit = str(release.get("source_commit") or "")
|
||||
version = str(release.get("version") or "")
|
||||
projection_id = release.get("projection_id")
|
||||
if CANDIDATE_RE.fullmatch(candidate_id) is None:
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_CANDIDATE_ID_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
if OID_RE.fullmatch(source_commit) is None:
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_SOURCE_COMMIT_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
if VERSION_RE.fullmatch(version) is None:
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_VERSION_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
if projection_id is not None:
|
||||
projection_id = str(projection_id)
|
||||
if not re.fullmatch(r"projection-[0-9a-f]{32}", projection_id):
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_PROJECTION_ID_INVALID",
|
||||
phase="validation",
|
||||
)
|
||||
return cls(
|
||||
candidate_id=candidate_id,
|
||||
source_commit=source_commit,
|
||||
version=version,
|
||||
projection_id=projection_id,
|
||||
)
|
||||
|
||||
def body(self) -> dict[str, str]:
|
||||
return {
|
||||
"candidate_id": self.candidate_id,
|
||||
"source_commit": self.source_commit,
|
||||
"version": self.version,
|
||||
}
|
||||
|
||||
|
||||
class Heartbeat:
|
||||
def __init__(
|
||||
self,
|
||||
agent_token: str,
|
||||
task_id: str,
|
||||
lease_token: str,
|
||||
) -> None:
|
||||
self.agent_token = agent_token
|
||||
self.task_id = task_id
|
||||
self.lease_token = lease_token
|
||||
self.stage = "claimed"
|
||||
self.failure: str | None = None
|
||||
self.stop = threading.Event()
|
||||
self.thread = threading.Thread(
|
||||
target=self._run,
|
||||
name="release-worker-heartbeat",
|
||||
daemon=True,
|
||||
)
|
||||
|
||||
def first(self) -> None:
|
||||
status, value = http_json(
|
||||
TASK_API,
|
||||
"POST",
|
||||
f"/v1/tasks/{self.task_id}/heartbeat",
|
||||
self.agent_token,
|
||||
{
|
||||
"lease_token": self.lease_token,
|
||||
"lease_seconds": LEASE_SECONDS,
|
||||
"progress": {
|
||||
"stage": self.stage,
|
||||
"worker_version": AGENT_VERSION,
|
||||
},
|
||||
},
|
||||
)
|
||||
if status != 200:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_HEARTBEAT_FAILED",
|
||||
phase="heartbeat",
|
||||
)
|
||||
audit(
|
||||
"TASK_HEARTBEAT",
|
||||
task_id=self.task_id,
|
||||
stage=self.stage,
|
||||
api_status=status,
|
||||
)
|
||||
write_liveness(
|
||||
"BUSY",
|
||||
task_id=self.task_id,
|
||||
stage=self.stage,
|
||||
)
|
||||
sd_notify("WATCHDOG=1")
|
||||
|
||||
def _run(self) -> None:
|
||||
while not self.stop.wait(HEARTBEAT_SECONDS):
|
||||
try:
|
||||
status, value = http_json(
|
||||
TASK_API,
|
||||
"POST",
|
||||
f"/v1/tasks/{self.task_id}/heartbeat",
|
||||
self.agent_token,
|
||||
{
|
||||
"lease_token": self.lease_token,
|
||||
"lease_seconds": LEASE_SECONDS,
|
||||
"progress": {
|
||||
"stage": self.stage,
|
||||
"worker_version": AGENT_VERSION,
|
||||
},
|
||||
},
|
||||
)
|
||||
if status != 200:
|
||||
self.failure = str(value.get("error") or status)
|
||||
return
|
||||
audit(
|
||||
"TASK_HEARTBEAT",
|
||||
task_id=self.task_id,
|
||||
stage=self.stage,
|
||||
api_status=status,
|
||||
)
|
||||
write_liveness(
|
||||
"BUSY",
|
||||
task_id=self.task_id,
|
||||
stage=self.stage,
|
||||
)
|
||||
sd_notify("WATCHDOG=1")
|
||||
except Exception as exc:
|
||||
self.failure = type(exc).__name__
|
||||
return
|
||||
|
||||
def ensure(self) -> None:
|
||||
if self.failure:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_LEASE_LOST",
|
||||
phase="heartbeat",
|
||||
)
|
||||
|
||||
def close(self) -> None:
|
||||
self.stop.set()
|
||||
if self.thread.ident is not None:
|
||||
self.thread.join(timeout=2)
|
||||
|
||||
|
||||
def release_admission(
|
||||
task: Mapping[str, Any],
|
||||
*,
|
||||
release_token: str,
|
||||
):
|
||||
required = task.get("required_capabilities")
|
||||
if not isinstance(required, list):
|
||||
required = []
|
||||
|
||||
def release_api_health() -> bool:
|
||||
status, value = http_json(
|
||||
RELEASE_API,
|
||||
"GET",
|
||||
"/health",
|
||||
"",
|
||||
timeout=10,
|
||||
)
|
||||
return status == 200 and str(value.get("status") or "").lower() in {"ok", "healthy"}
|
||||
|
||||
def authenticated_probe() -> bool:
|
||||
status, _ = http_json(
|
||||
RELEASE_API,
|
||||
"POST",
|
||||
"/createReleaseCandidate",
|
||||
release_token,
|
||||
{},
|
||||
timeout=10,
|
||||
)
|
||||
# Authentication + credential projection succeeded; the empty release
|
||||
# request is intentionally rejected by request validation.
|
||||
return status == 400
|
||||
|
||||
probes = {
|
||||
"credential_projection": authenticated_probe,
|
||||
"trusted_builder": release_api_health,
|
||||
"authenticated_probe": authenticated_probe,
|
||||
"release_api_health": release_api_health,
|
||||
"worker_liveness": lambda: True,
|
||||
}
|
||||
return runtime_admission(
|
||||
worker_capabilities=EXPECTED_CAPABILITIES,
|
||||
required_capabilities=required,
|
||||
requested_operations=("prepare_release",),
|
||||
probes=probes,
|
||||
)
|
||||
|
||||
|
||||
def result_payload(
|
||||
task_id: str,
|
||||
release: ReleaseRequest,
|
||||
builder: Mapping[str, Any],
|
||||
) -> dict[str, Any]:
|
||||
builder_candidate = builder.get("candidate_id")
|
||||
if builder_candidate not in {None, release.candidate_id}:
|
||||
raise ReleaseWorkerError(
|
||||
"BUILDER_CANDIDATE_BINDING_INVALID",
|
||||
phase="release",
|
||||
)
|
||||
immutable = {
|
||||
"candidate_id": release.candidate_id,
|
||||
"source_commit": release.source_commit,
|
||||
"version": release.version,
|
||||
}
|
||||
return {
|
||||
"status": "SUCCEEDED",
|
||||
"classification": "RELEASE_CANDIDATE_CREATED",
|
||||
"execution_mode": "release_candidate",
|
||||
"adapter": "release_candidate",
|
||||
"task_id": task_id,
|
||||
"agent_id": AGENT_ID,
|
||||
"release_candidate": immutable,
|
||||
"release_candidate_request_sha256": sha256(canonical(immutable)),
|
||||
"builder_response": sanitized(dict(builder)),
|
||||
"projection_id": release.projection_id,
|
||||
"writer_lease": "NONE",
|
||||
"changed_files": [],
|
||||
"direct_main_write": False,
|
||||
"production_write": False,
|
||||
"tag_write": False,
|
||||
"baseline_write": False,
|
||||
"runtime_switch": False,
|
||||
"owner_approval_submitted": False,
|
||||
"credential_material_exposed": False,
|
||||
}
|
||||
|
||||
|
||||
def execute_task(
|
||||
task: Mapping[str, Any],
|
||||
*,
|
||||
agent_token: str,
|
||||
release_token: str,
|
||||
) -> bool:
|
||||
task_id = str(task.get("task_id") or "")
|
||||
release = ReleaseRequest.from_task(task)
|
||||
admission = release_admission(
|
||||
task,
|
||||
release_token=release_token,
|
||||
)
|
||||
if not admission.admitted:
|
||||
audit(
|
||||
"TASK_ADMISSION_BLOCKED",
|
||||
task_id=task_id,
|
||||
admission=admission.as_dict(),
|
||||
)
|
||||
write_liveness(
|
||||
"BLOCKED",
|
||||
task_id=task_id,
|
||||
stage="execution_ready",
|
||||
error=admission.failure_class.value if admission.failure_class else "ADMISSION_FAILED",
|
||||
)
|
||||
return False
|
||||
audit(
|
||||
"TASK_ADMISSION_PASSED",
|
||||
task_id=task_id,
|
||||
admission=admission.as_dict(),
|
||||
)
|
||||
status, claimed = http_json(
|
||||
TASK_API,
|
||||
"POST",
|
||||
f"/v1/tasks/{task_id}/claim",
|
||||
agent_token,
|
||||
{"lease_seconds": LEASE_SECONDS},
|
||||
)
|
||||
if status in {409, 423}:
|
||||
return False
|
||||
if status != 200:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_CLAIM_FAILED",
|
||||
phase="claim",
|
||||
)
|
||||
lease_token = str(claimed.get("lease_token") or "")
|
||||
if len(lease_token) < 16:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_LEASE_TOKEN_INVALID",
|
||||
phase="claim",
|
||||
)
|
||||
attempt = int(claimed.get("attempt") or 0)
|
||||
audit(
|
||||
"TASK_CLAIMED",
|
||||
task_id=task_id,
|
||||
attempt=attempt,
|
||||
)
|
||||
|
||||
heartbeat = Heartbeat(agent_token, task_id, lease_token)
|
||||
heartbeat.first()
|
||||
heartbeat.thread.start()
|
||||
try:
|
||||
heartbeat.stage = "release_candidate"
|
||||
heartbeat.ensure()
|
||||
status, builder = http_json(
|
||||
RELEASE_API,
|
||||
"POST",
|
||||
"/createReleaseCandidate",
|
||||
release_token,
|
||||
release.body(),
|
||||
timeout=120,
|
||||
)
|
||||
if status not in {200, 201, 202}:
|
||||
raise ReleaseWorkerError(
|
||||
"RELEASE_CANDIDATE_CREATE_FAILED:"
|
||||
+ str(builder.get("error") or status),
|
||||
phase="release",
|
||||
)
|
||||
heartbeat.ensure()
|
||||
result = result_payload(task_id, release, builder)
|
||||
idempotency_key = (
|
||||
f"{task_id}-attempt-{attempt}-release-result-v1"
|
||||
)
|
||||
status, submitted = http_json(
|
||||
TASK_API,
|
||||
"POST",
|
||||
f"/v1/tasks/{task_id}/result",
|
||||
agent_token,
|
||||
{
|
||||
"lease_token": lease_token,
|
||||
"idempotency_key": idempotency_key,
|
||||
"result": result,
|
||||
},
|
||||
)
|
||||
if status not in {200, 201, 202}:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_RESULT_SUBMIT_FAILED",
|
||||
phase="result",
|
||||
)
|
||||
audit(
|
||||
"RESULT_SUBMITTED",
|
||||
task_id=task_id,
|
||||
attempt=attempt,
|
||||
release_candidate_id=release.candidate_id,
|
||||
)
|
||||
return True
|
||||
finally:
|
||||
heartbeat.close()
|
||||
|
||||
|
||||
def main() -> int:
|
||||
if not RELEASE_API:
|
||||
raise SystemExit("FATAL:AIOS_RELEASE_API_URL_REQUIRED")
|
||||
agent_token = load_credential("agent_token")
|
||||
release_token = load_credential("release_candidate_token")
|
||||
STATE_ROOT.mkdir(parents=True, exist_ok=True, mode=0o750)
|
||||
audit(
|
||||
"WORKER_STARTED",
|
||||
capabilities=sorted(EXPECTED_CAPABILITIES),
|
||||
release_api=RELEASE_API,
|
||||
)
|
||||
write_liveness("STARTING", stage="startup")
|
||||
sd_notify("READY=1\nSTATUS=release worker ready")
|
||||
while True:
|
||||
try:
|
||||
status, response = http_json(
|
||||
TASK_API,
|
||||
"POST",
|
||||
"/v1/tasks/poll",
|
||||
agent_token,
|
||||
{"limit": 100},
|
||||
)
|
||||
if status != 200:
|
||||
raise ReleaseWorkerError(
|
||||
"TASK_POLL_FAILED",
|
||||
phase="poll",
|
||||
)
|
||||
tasks = response.get("tasks")
|
||||
if not isinstance(tasks, list):
|
||||
tasks = []
|
||||
write_liveness(
|
||||
"IDLE",
|
||||
stage="poll",
|
||||
visible_count=len(tasks),
|
||||
)
|
||||
sd_notify("WATCHDOG=1\nSTATUS=release worker poll ok")
|
||||
handled = False
|
||||
for raw in tasks:
|
||||
if not isinstance(raw, Mapping):
|
||||
continue
|
||||
caps = raw.get("required_capabilities")
|
||||
if (
|
||||
not isinstance(caps, list)
|
||||
or frozenset(caps) != EXPECTED_CAPABILITIES
|
||||
):
|
||||
continue
|
||||
try:
|
||||
handled = execute_task(
|
||||
raw,
|
||||
agent_token=agent_token,
|
||||
release_token=release_token,
|
||||
) or handled
|
||||
except ReleaseWorkerError as exc:
|
||||
failure = classify_failure(exc.code, phase=exc.phase)
|
||||
repair = repair_for_failure(failure)
|
||||
audit(
|
||||
"TASK_EXECUTION_FAILED",
|
||||
task_id=raw.get("task_id"),
|
||||
phase=exc.phase,
|
||||
error=exc.code,
|
||||
failure_class=failure.value,
|
||||
repair=repair.as_dict(),
|
||||
)
|
||||
time.sleep(0 if handled else POLL_SECONDS)
|
||||
except KeyboardInterrupt:
|
||||
return 0
|
||||
except Exception as exc:
|
||||
audit(
|
||||
"WORKER_LOOP_ERROR",
|
||||
error=type(exc).__name__,
|
||||
)
|
||||
write_liveness(
|
||||
"DEGRADED",
|
||||
stage="loop",
|
||||
error=type(exc).__name__,
|
||||
)
|
||||
time.sleep(POLL_SECONDS)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -0,0 +1,287 @@
|
||||
"""Credential-isolated durable coordinator for the release-preparation supervisor.
|
||||
|
||||
The coordinator owns only its Online Task agent credential. It never forwards
|
||||
that credential to Hermes/OpenClaw and never submits a Task result while the
|
||||
supervision goal is incomplete. Its job is deliberately small: claim the one
|
||||
fixed supervisor Task after an exact coordinator projection and keep a durable
|
||||
heartbeat lease while later phases observe and repair through governed APIs.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime, timezone
|
||||
import json
|
||||
import os
|
||||
from pathlib import Path
|
||||
import tempfile
|
||||
import time
|
||||
from typing import Any, Mapping
|
||||
import urllib.error
|
||||
import urllib.request
|
||||
|
||||
|
||||
API_BASE = os.environ.get(
|
||||
"AIOS_AGENT_TASK_API_URL", "http://127.0.0.1:18940"
|
||||
).rstrip("/")
|
||||
AGENT_ID = os.environ.get(
|
||||
"AIOS_AGENT_ID",
|
||||
"online-engineering-worker-cn-network-mkt-autonomous-20260815",
|
||||
)
|
||||
TASK_ID = os.environ.get(
|
||||
"AIOS_RELEASE_PREPARATION_SUPERVISOR_TASK_ID",
|
||||
"r42-release-preparation-autonomous-supervisor-20260820",
|
||||
)
|
||||
EXPECTED_EXECUTION_DIGEST = os.environ.get(
|
||||
"AIOS_RELEASE_PREPARATION_SUPERVISOR_EXECUTION_DIGEST", ""
|
||||
)
|
||||
RUNTIME_CAPABILITY = "agent.runtime.openclaw"
|
||||
COORDINATOR_CAPABILITY = "engineering.autonomous"
|
||||
REQUIRED_CAPABILITIES = [RUNTIME_CAPABILITY, COORDINATOR_CAPABILITY]
|
||||
POLL_SECONDS = int(os.environ.get("AIOS_SUPERVISOR_COORDINATOR_POLL_SECONDS", "5"))
|
||||
HEARTBEAT_SECONDS = int(
|
||||
os.environ.get("AIOS_SUPERVISOR_COORDINATOR_HEARTBEAT_SECONDS", "10")
|
||||
)
|
||||
LEASE_SECONDS = int(os.environ.get("AIOS_SUPERVISOR_COORDINATOR_LEASE_SECONDS", "300"))
|
||||
STATE_ROOT = Path(
|
||||
os.environ.get(
|
||||
"AIOS_SUPERVISOR_COORDINATOR_STATE_ROOT",
|
||||
"/var/lib/aios-release-preparation-supervisor-coordinator",
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
class CoordinatorError(RuntimeError):
|
||||
def __init__(self, code: str) -> None:
|
||||
super().__init__(code)
|
||||
self.code = code
|
||||
|
||||
|
||||
def now() -> str:
|
||||
return datetime.now(timezone.utc).isoformat()
|
||||
|
||||
|
||||
def canonical(value: object) -> bytes:
|
||||
return json.dumps(
|
||||
value, sort_keys=True, separators=(",", ":"), ensure_ascii=False
|
||||
).encode("utf-8")
|
||||
|
||||
|
||||
def atomic_json(path: Path, value: Mapping[str, Any]) -> None:
|
||||
path.parent.mkdir(parents=True, mode=0o750, exist_ok=True)
|
||||
if path.parent.is_symlink() or path.is_symlink():
|
||||
raise CoordinatorError("COORDINATOR_STATE_SYMLINK_REJECTED")
|
||||
fd, temporary = tempfile.mkstemp(prefix=f".{path.name}.", dir=str(path.parent))
|
||||
try:
|
||||
os.fchmod(fd, 0o640)
|
||||
with os.fdopen(fd, "w", encoding="utf-8") as handle:
|
||||
json.dump(dict(value), handle, ensure_ascii=False, sort_keys=True)
|
||||
handle.write("\n")
|
||||
handle.flush()
|
||||
os.fsync(handle.fileno())
|
||||
os.replace(temporary, path)
|
||||
finally:
|
||||
if os.path.exists(temporary):
|
||||
os.unlink(temporary)
|
||||
|
||||
|
||||
def audit(event_type: str, **fields: Any) -> None:
|
||||
record = {
|
||||
"event_type": event_type,
|
||||
"created_at": now(),
|
||||
"agent_id": AGENT_ID,
|
||||
"task_id": TASK_ID,
|
||||
"credential_material_exposed": False,
|
||||
"planner_credential_material_exposed": False,
|
||||
"lease_material_exposed": False,
|
||||
**fields,
|
||||
}
|
||||
STATE_ROOT.mkdir(parents=True, mode=0o750, exist_ok=True)
|
||||
with (STATE_ROOT / "events.jsonl").open("a", encoding="utf-8") as handle:
|
||||
handle.write(json.dumps(record, ensure_ascii=False, sort_keys=True) + "\n")
|
||||
atomic_json(STATE_ROOT / "liveness.json", record)
|
||||
|
||||
|
||||
def api_call(
|
||||
method: str,
|
||||
route: str,
|
||||
token: str,
|
||||
body: Mapping[str, Any] | None = None,
|
||||
) -> tuple[int, dict[str, Any]]:
|
||||
payload = None if body is None else canonical(dict(body))
|
||||
headers = {"Authorization": "Bearer " + token, "Accept": "application/json"}
|
||||
if payload is not None:
|
||||
headers["Content-Type"] = "application/json"
|
||||
request = urllib.request.Request(
|
||||
API_BASE + route, data=payload, headers=headers, method=method
|
||||
)
|
||||
try:
|
||||
with urllib.request.urlopen(request, timeout=20) as response:
|
||||
value = json.loads(response.read().decode("utf-8"))
|
||||
return response.status, value if isinstance(value, dict) else {}
|
||||
except urllib.error.HTTPError as exc:
|
||||
try:
|
||||
value = json.loads(exc.read(4096).decode("utf-8", "replace"))
|
||||
except Exception:
|
||||
value = {"error": "HTTP_ERROR"}
|
||||
return exc.code, value if isinstance(value, dict) else {"error": "HTTP_ERROR"}
|
||||
except (urllib.error.URLError, TimeoutError):
|
||||
return 599, {"error": "AGENT_TASK_API_UNAVAILABLE"}
|
||||
|
||||
|
||||
def load_agent_token() -> str:
|
||||
path = Path(os.environ.get("CREDENTIALS_DIRECTORY", "")) / "agent_token"
|
||||
if path.is_symlink() or not path.is_file():
|
||||
raise CoordinatorError("COORDINATOR_AGENT_CREDENTIAL_UNAVAILABLE")
|
||||
token = path.read_text(encoding="utf-8").strip()
|
||||
if len(token) < 24 or any(character.isspace() for character in token):
|
||||
raise CoordinatorError("COORDINATOR_AGENT_CREDENTIAL_INVALID")
|
||||
return token
|
||||
|
||||
|
||||
def validate_task(task: Mapping[str, Any]) -> dict[str, Any]:
|
||||
if task.get("task_id") != TASK_ID or task.get("status") != "QUEUED":
|
||||
raise CoordinatorError("SUPERVISOR_TASK_STATE_INVALID")
|
||||
if task.get("required_capabilities") != REQUIRED_CAPABILITIES:
|
||||
raise CoordinatorError("SUPERVISOR_COORDINATOR_CAPABILITIES_INVALID")
|
||||
metadata = task.get("metadata")
|
||||
if not isinstance(metadata, Mapping):
|
||||
raise CoordinatorError("SUPERVISOR_EXECUTION_METADATA_MISSING")
|
||||
execution = metadata.get("execution")
|
||||
if not isinstance(execution, Mapping):
|
||||
raise CoordinatorError("SUPERVISOR_EXECUTION_METADATA_MISSING")
|
||||
governance = execution.get("governance")
|
||||
if (
|
||||
execution.get("action") != "agent_runtime"
|
||||
or execution.get("adapter") != "openclaw"
|
||||
or execution.get("capability_scope") != REQUIRED_CAPABILITIES
|
||||
or execution.get("execution_digest") != EXPECTED_EXECUTION_DIGEST
|
||||
or execution.get("change_set") != []
|
||||
or execution.get("verification") != []
|
||||
or not isinstance(governance, Mapping)
|
||||
or governance.get("writer_lease") is not False
|
||||
or any(
|
||||
governance.get(field) is not False
|
||||
for field in (
|
||||
"main_write",
|
||||
"production_write",
|
||||
"tag_write",
|
||||
"baseline_write",
|
||||
"force_push",
|
||||
"protected_ref_bypass",
|
||||
)
|
||||
)
|
||||
):
|
||||
raise CoordinatorError("SUPERVISOR_COORDINATOR_ENVELOPE_INVALID")
|
||||
return dict(execution)
|
||||
|
||||
|
||||
def heartbeat_progress(sequence: int) -> dict[str, Any]:
|
||||
return {
|
||||
"stage": "supervisor_coordinator_active",
|
||||
"phase": "external_caller_projection",
|
||||
"state": "PENDING_EXTERNAL_CONNECTOR_ACK",
|
||||
"sequence": sequence,
|
||||
"control_plane_access_granted": True,
|
||||
"external_planner_role": "UNPRIVILEGED_PLANNER",
|
||||
"credential_forwarded_to_planner": False,
|
||||
"main_write": False,
|
||||
"production_write": False,
|
||||
"owner_approval_submitted": False,
|
||||
}
|
||||
|
||||
|
||||
def coordinate_claim(
|
||||
claim: Mapping[str, Any],
|
||||
token: str,
|
||||
*,
|
||||
max_heartbeats: int | None = None,
|
||||
) -> int:
|
||||
lease_token = claim.get("lease_token")
|
||||
if not isinstance(lease_token, str) or len(lease_token) < 24:
|
||||
raise CoordinatorError("COORDINATOR_LEASE_TOKEN_MISSING")
|
||||
sequence = 0
|
||||
while max_heartbeats is None or sequence < max_heartbeats:
|
||||
sequence += 1
|
||||
status, response = api_call(
|
||||
"POST",
|
||||
f"/v1/tasks/{TASK_ID}/heartbeat",
|
||||
token,
|
||||
{
|
||||
"lease_token": lease_token,
|
||||
"lease_seconds": LEASE_SECONDS,
|
||||
"progress": heartbeat_progress(sequence),
|
||||
},
|
||||
)
|
||||
if status != 200:
|
||||
raise CoordinatorError(
|
||||
"COORDINATOR_HEARTBEAT_FAILED_" + str(response.get("error") or status)
|
||||
)
|
||||
audit(
|
||||
"SUPERVISOR_COORDINATOR_HEARTBEAT",
|
||||
sequence=sequence,
|
||||
task_status=response.get("status"),
|
||||
lease_expires_at=response.get("lease_expires_at"),
|
||||
control_plane_access_granted=True,
|
||||
external_planner_role="UNPRIVILEGED_PLANNER",
|
||||
)
|
||||
if max_heartbeats is None or sequence < max_heartbeats:
|
||||
time.sleep(HEARTBEAT_SECONDS)
|
||||
return sequence
|
||||
|
||||
|
||||
def worker_once(token: str) -> bool:
|
||||
status, poll = api_call("POST", "/v1/tasks/poll", token, {"limit": 100})
|
||||
if status != 200:
|
||||
audit("SUPERVISOR_COORDINATOR_POLL_FAILED", error=poll.get("error") or status)
|
||||
return False
|
||||
task = next(
|
||||
(
|
||||
item
|
||||
for item in poll.get("tasks", [])
|
||||
if isinstance(item, Mapping) and item.get("task_id") == TASK_ID
|
||||
),
|
||||
None,
|
||||
)
|
||||
if task is None:
|
||||
audit("SUPERVISOR_COORDINATOR_WAITING_FOR_TASK")
|
||||
return False
|
||||
validate_task(task)
|
||||
status, claim = api_call(
|
||||
"POST", f"/v1/tasks/{TASK_ID}/claim", token, {"lease_seconds": LEASE_SECONDS}
|
||||
)
|
||||
if status != 200:
|
||||
audit("SUPERVISOR_COORDINATOR_CLAIM_FAILED", error=claim.get("error") or status)
|
||||
return False
|
||||
audit(
|
||||
"SUPERVISOR_COORDINATOR_CLAIMED",
|
||||
attempt=claim.get("attempt"),
|
||||
revision=claim.get("revision"),
|
||||
claim_mode="server_atomic",
|
||||
control_plane_access_granted=True,
|
||||
)
|
||||
coordinate_claim(claim, token)
|
||||
return True
|
||||
|
||||
|
||||
def main() -> int:
|
||||
if not EXPECTED_EXECUTION_DIGEST or len(EXPECTED_EXECUTION_DIGEST) != 64:
|
||||
raise CoordinatorError("SUPERVISOR_EXECUTION_DIGEST_CONFIG_INVALID")
|
||||
token = load_agent_token()
|
||||
audit(
|
||||
"SUPERVISOR_COORDINATOR_READY",
|
||||
required_capabilities=REQUIRED_CAPABILITIES,
|
||||
control_plane_access_granted=True,
|
||||
external_planner_role="UNPRIVILEGED_PLANNER",
|
||||
)
|
||||
while True:
|
||||
try:
|
||||
worker_once(token)
|
||||
except Exception as exc:
|
||||
audit(
|
||||
"SUPERVISOR_COORDINATOR_LOOP_ERROR",
|
||||
error=exc.code if isinstance(exc, CoordinatorError) else type(exc).__name__,
|
||||
)
|
||||
time.sleep(POLL_SECONDS)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -0,0 +1,37 @@
|
||||
[Unit]
|
||||
Description=AIOS owner-controlled Task initial execution projection API
|
||||
After=aios-agent-task-api.service aios-chatgpt-bridge-v2-staging.service
|
||||
Wants=aios-agent-task-api.service aios-chatgpt-bridge-v2-staging.service
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
User=aios-agent-task
|
||||
Group=aios-agent-task
|
||||
LoadCredential=agent_task_control_token:/etc/aios-gitops/agent-task-control.token
|
||||
Environment=AIOS_AGENT_TASK_DB=/var/lib/aios-agent-task/state.sqlite3
|
||||
Environment=AIOS_AGENT_TASK_CONTROL_TOKEN_FILE=%d/agent_task_control_token
|
||||
Environment=AIOS_GOVERNED_WRITE_API_URL=http://127.0.0.1:18892
|
||||
Environment=AIOS_AGENT_TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZATION_ID=owner-initial-execution-projection-recovery-20260820
|
||||
Environment=AIOS_TASK_INITIAL_EXECUTION_PROJECTION_HOST=127.0.0.1
|
||||
Environment=AIOS_TASK_INITIAL_EXECUTION_PROJECTION_PORT=18943
|
||||
ExecStart=/usr/bin/python3 -B /opt/aios/current/active-runtime/app/aios_agent_task_initial_execution_projection.py
|
||||
Restart=on-failure
|
||||
RestartSec=2
|
||||
NoNewPrivileges=true
|
||||
PrivateTmp=true
|
||||
ProtectSystem=strict
|
||||
ProtectHome=true
|
||||
ProtectKernelTunables=true
|
||||
ProtectKernelModules=true
|
||||
ProtectControlGroups=true
|
||||
LockPersonality=true
|
||||
RestrictSUIDSGID=true
|
||||
RestrictRealtime=true
|
||||
RestrictNamespaces=true
|
||||
MemoryDenyWriteExecute=true
|
||||
RestrictAddressFamilies=AF_UNIX AF_INET
|
||||
ReadWritePaths=/var/lib/aios-agent-task
|
||||
UMask=0077
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,37 @@
|
||||
[Unit]
|
||||
Description=AIOS owner-controlled supervisor coordinator projection API
|
||||
After=aios-agent-task-api.service aios-chatgpt-bridge-v2-staging.service
|
||||
Wants=aios-agent-task-api.service aios-chatgpt-bridge-v2-staging.service
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
User=aios-agent-task
|
||||
Group=aios-agent-task
|
||||
LoadCredential=agent_task_control_token:/etc/aios-gitops/agent-task-control.token
|
||||
Environment=AIOS_AGENT_TASK_DB=/var/lib/aios-agent-task/state.sqlite3
|
||||
Environment=AIOS_AGENT_TASK_CONTROL_TOKEN_FILE=%d/agent_task_control_token
|
||||
Environment=AIOS_GOVERNED_WRITE_API_URL=http://127.0.0.1:18892
|
||||
Environment=AIOS_AGENT_TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZATION_ID=owner-supervisor-coordinator-projection-recovery-20260820
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_PROJECTION_HOST=127.0.0.1
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_PROJECTION_PORT=18944
|
||||
ExecStart=/usr/bin/python3 -B /opt/aios/current/active-runtime/app/aios_agent_task_supervisor_coordinator_projection.py
|
||||
Restart=on-failure
|
||||
RestartSec=2
|
||||
NoNewPrivileges=true
|
||||
PrivateTmp=true
|
||||
ProtectSystem=strict
|
||||
ProtectHome=true
|
||||
ProtectKernelTunables=true
|
||||
ProtectKernelModules=true
|
||||
ProtectControlGroups=true
|
||||
LockPersonality=true
|
||||
RestrictSUIDSGID=true
|
||||
RestrictRealtime=true
|
||||
RestrictNamespaces=true
|
||||
MemoryDenyWriteExecute=true
|
||||
RestrictAddressFamilies=AF_UNIX AF_INET
|
||||
ReadWritePaths=/var/lib/aios-agent-task
|
||||
UMask=0077
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,47 @@
|
||||
[Unit]
|
||||
Description=AIOS durable Owner authorization broker
|
||||
After=network.target aios-chatgpt-bridge-v2-staging.service
|
||||
Wants=aios-chatgpt-bridge-v2-staging.service
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
User=root
|
||||
Group=aios-executor-exchange
|
||||
RuntimeDirectory=aios-owner-authorization-broker
|
||||
RuntimeDirectoryMode=0750
|
||||
StateDirectory=aios-owner-authorization-broker
|
||||
StateDirectoryMode=0750
|
||||
LoadCredential=agent_task_control_token:/etc/aios-gitops/agent-task-control.token
|
||||
Environment=AIOS_AGENT_TASK_CONTROL_TOKEN_FILE=%d/agent_task_control_token
|
||||
Environment=AIOS_OWNER_DURABLE_AUTHORIZATION_POLICY=/etc/aios-gitops/owner-durable-authorization.json
|
||||
Environment=AIOS_OWNER_DURABLE_AUTHORIZATION_REVOCATIONS=/etc/aios-gitops/owner-durable-authorization-revocations.json
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_ACCOUNT=chatgpt-governed-implementer-singlehost
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_SOCKET=%t/aios-owner-authorization-broker/broker.sock
|
||||
ExecStart=/usr/bin/python3 -B /opt/aios/current/active-runtime/app/aios_durable_owner_authorization.py
|
||||
Restart=on-failure
|
||||
RestartSec=2
|
||||
NoNewPrivileges=yes
|
||||
PrivateTmp=yes
|
||||
PrivateDevices=yes
|
||||
ProtectSystem=strict
|
||||
ProtectHome=yes
|
||||
ProtectKernelTunables=yes
|
||||
ProtectKernelModules=yes
|
||||
ProtectControlGroups=yes
|
||||
ProtectKernelLogs=yes
|
||||
ProtectClock=yes
|
||||
LockPersonality=yes
|
||||
RestrictSUIDSGID=yes
|
||||
RestrictRealtime=yes
|
||||
RestrictNamespaces=yes
|
||||
MemoryDenyWriteExecute=yes
|
||||
RestrictAddressFamilies=AF_UNIX AF_INET
|
||||
IPAddressDeny=any
|
||||
IPAddressAllow=127.0.0.0/8
|
||||
ReadOnlyPaths=/etc/aios-gitops/owner-durable-authorization.json
|
||||
ReadOnlyPaths=/etc/aios-gitops/owner-durable-authorization-revocations.json
|
||||
InaccessiblePaths=/data/aios-gitops /home
|
||||
UMask=0007
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,50 @@
|
||||
[Unit]
|
||||
Description=AIOS release-candidate Online Agent worker
|
||||
After=network.target aios-agent-task-api.service
|
||||
Wants=aios-agent-task-api.service
|
||||
|
||||
[Service]
|
||||
Type=notify
|
||||
NotifyAccess=main
|
||||
WatchdogSec=90
|
||||
User=root
|
||||
Group=root
|
||||
StateDirectory=aios-online-agent-release
|
||||
StateDirectoryMode=0750
|
||||
LoadCredentialEncrypted=agent_token:/etc/credstore.encrypted/aios-release-online-agent-worker/agent_token
|
||||
LoadCredentialEncrypted=release_candidate_token:/etc/credstore.encrypted/aios-release-online-agent-worker/release_candidate_token
|
||||
Environment=AIOS_RELEASE_WORKER_AGENT_ID=online-release-worker-cn-network-mkt-20260819
|
||||
Environment=AIOS_RELEASE_WORKER_AGENT_VERSION=2026-08-19.1-release-candidate
|
||||
Environment=AIOS_AGENT_TASK_API_URL=http://127.0.0.1:18940
|
||||
Environment=AIOS_RELEASE_WORKER_STATE_ROOT=/var/lib/aios-online-agent-release
|
||||
Environment=AIOS_WORKER_STATE_ROOT=/var/lib/aios-online-agent-release
|
||||
Environment=AIOS_RELEASE_WORKER_POLL_SECONDS=3
|
||||
Environment=AIOS_RELEASE_WORKER_LEASE_SECONDS=60
|
||||
Environment=AIOS_RELEASE_WORKER_HEARTBEAT_SECONDS=15
|
||||
Environment=AIOS_RELEASE_API_URL=http://127.0.0.1:18890
|
||||
ExecStart=/usr/bin/python3 -B /opt/aios/current/active-runtime/app/aios_release_online_agent_worker.py
|
||||
Restart=always
|
||||
RestartSec=2
|
||||
NoNewPrivileges=yes
|
||||
PrivateTmp=yes
|
||||
PrivateDevices=yes
|
||||
ProtectSystem=strict
|
||||
ProtectHome=yes
|
||||
ProtectKernelTunables=yes
|
||||
ProtectKernelModules=yes
|
||||
ProtectControlGroups=yes
|
||||
ProtectKernelLogs=yes
|
||||
ProtectClock=yes
|
||||
LockPersonality=yes
|
||||
RestrictSUIDSGID=yes
|
||||
RestrictRealtime=yes
|
||||
RestrictNamespaces=yes
|
||||
MemoryDenyWriteExecute=yes
|
||||
RestrictAddressFamilies=AF_UNIX AF_INET
|
||||
IPAddressDeny=any
|
||||
IPAddressAllow=127.0.0.0/8
|
||||
InaccessiblePaths=/data/aios-gitops /home
|
||||
UMask=0077
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,45 @@
|
||||
[Unit]
|
||||
Description=AIOS credential-isolated release-preparation supervisor coordinator
|
||||
After=network-online.target aios-agent-task-api.service aios-autonomous-agent-register.service
|
||||
Wants=network-online.target aios-agent-task-api.service
|
||||
Requires=aios-autonomous-agent-register.service
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
User=aios-autonomous-agent-worker
|
||||
Group=aios-autonomous-agent-worker
|
||||
SupplementaryGroups=aios-agent-runtime
|
||||
LoadCredentialEncrypted=agent_token:/etc/credstore.encrypted/aios-autonomous-agent-worker/agent_token
|
||||
Environment=PYTHONUNBUFFERED=1
|
||||
Environment=AIOS_AGENT_TASK_API_URL=http://127.0.0.1:18940
|
||||
Environment=AIOS_AGENT_ID=online-engineering-worker-cn-network-mkt-autonomous-20260815
|
||||
Environment=AIOS_RELEASE_PREPARATION_SUPERVISOR_TASK_ID=r42-release-preparation-autonomous-supervisor-20260820
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_STATE_ROOT=/var/lib/aios-release-preparation-supervisor-coordinator
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_POLL_SECONDS=5
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_HEARTBEAT_SECONDS=10
|
||||
Environment=AIOS_SUPERVISOR_COORDINATOR_LEASE_SECONDS=300
|
||||
ExecStart=/usr/bin/python3 -B /opt/aios/current/active-runtime/app/aios_release_preparation_supervisor_coordinator.py
|
||||
Restart=always
|
||||
RestartSec=5
|
||||
TimeoutStopSec=20
|
||||
UMask=0027
|
||||
NoNewPrivileges=true
|
||||
PrivateTmp=true
|
||||
ProtectSystem=strict
|
||||
ProtectHome=true
|
||||
ProtectKernelTunables=true
|
||||
ProtectKernelModules=true
|
||||
ProtectControlGroups=true
|
||||
LockPersonality=true
|
||||
RestrictSUIDSGID=true
|
||||
RestrictRealtime=true
|
||||
RestrictNamespaces=true
|
||||
MemoryDenyWriteExecute=true
|
||||
RestrictAddressFamilies=AF_INET AF_UNIX
|
||||
ReadWritePaths=/var/lib/aios-release-preparation-supervisor-coordinator
|
||||
InaccessiblePaths=-/etc/credstore.encrypted
|
||||
CapabilityBoundingSet=
|
||||
AmbientCapabilities=
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -160,6 +160,23 @@ components:
|
||||
properties:
|
||||
text: {type: string}
|
||||
branch: {type: string}
|
||||
CreateGovernedTaskRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [task_id, repository, base_oid, target_branch, risk, description, idempotency_key]
|
||||
properties:
|
||||
task_id: {type: string, minLength: 1}
|
||||
repository: {type: string, const: aios/aios}
|
||||
base_oid: {type: string, pattern: '^[0-9a-f]{40}$'}
|
||||
target_branch: {type: string, pattern: '^agent/[A-Za-z0-9][A-Za-z0-9._/-]{2,179}$'}
|
||||
risk: {type: string, enum: [low, medium, high, engineering]}
|
||||
description: {type: string, minLength: 1}
|
||||
idempotency_key: {type: string, minLength: 1, maxLength: 256}
|
||||
expected_branch_oid: {type: string, pattern: '^[0-9a-f]{40}$'}
|
||||
generation: {type: integer, minimum: 1}
|
||||
parent_task_id: {type: string, minLength: 1}
|
||||
parent_job_id: {type: string, pattern: '^job-[A-Za-z0-9][A-Za-z0-9._-]{7,255}$'}
|
||||
priority: {type: integer, minimum: -100, maximum: 100}
|
||||
DeploymentPlan:
|
||||
type: object
|
||||
required: [service, release_path, release_root, cutover_link, dropin_path, dropin_sha256, systemd_unit_sha256, rollback_path, rollback_sha256, source_commit, tree_oid, artifact_sha256, manifest_sha256, package_closure_sha256, authoritative_git_dir, repository_identity, writer_expected, baseline_before_sha256, registry_before_sha256, production_tag_refset_before_sha256, health_contract_version, timeout_policy, retry_policy, policy_sha256, health_contract]
|
||||
@@ -349,6 +366,23 @@ components:
|
||||
operations: {type: array, items: {type: string}}
|
||||
operation_contracts: {type: object, additionalProperties: true}
|
||||
paths:
|
||||
/createGovernedTask:
|
||||
post:
|
||||
operationId: createGovernedTask
|
||||
summary: Create or idempotently return a governed task binding through the existing Bridge
|
||||
security: [{bearerAuth: []}]
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema: {$ref: '#/components/schemas/CreateGovernedTaskRequest'}
|
||||
responses:
|
||||
'200': {description: Governed task binding created or returned idempotently}
|
||||
'400': {description: Invalid governed task or continuation binding}
|
||||
'401': {description: Caller authorization required}
|
||||
'409': {description: Durable task, parent, branch, or generation conflict}
|
||||
'503': {description: Governed Bridge or server-side control credential unavailable}
|
||||
x-openai-isConsequential: true
|
||||
/createReleaseIntent:
|
||||
post:
|
||||
operationId: createReleaseIntent
|
||||
|
||||
@@ -0,0 +1,210 @@
|
||||
# AIOS Autonomous Control Plane — Admission + Invariant + Repair
|
||||
|
||||
## Why this exists
|
||||
|
||||
Recent Agent registration, Task claim, release-worker, credential and retry failures shared one
|
||||
failure mode:
|
||||
|
||||
> The control plane accepted an intent before the runtime had proved it could execute the intent.
|
||||
|
||||
This design does not add another fleet of daemons. It adds one shared library contract that Task API,
|
||||
workers and governed host recovery consume.
|
||||
|
||||
## Three layers
|
||||
|
||||
```text
|
||||
Intent
|
||||
↓
|
||||
Runtime Admission
|
||||
↓
|
||||
System Invariants
|
||||
↓
|
||||
Repair Policy
|
||||
↓
|
||||
Execution
|
||||
```
|
||||
|
||||
### Runtime Admission
|
||||
|
||||
Admission occurs before Task claim.
|
||||
|
||||
A capability such as `release:candidate` is not considered executable merely because its name
|
||||
matches. The capability contract also binds:
|
||||
|
||||
- supported contract version;
|
||||
- allowed permission;
|
||||
- forbidden operations;
|
||||
- runtime dependencies;
|
||||
- health requirements.
|
||||
|
||||
For `release:candidate`:
|
||||
|
||||
```text
|
||||
permission:
|
||||
prepare_release
|
||||
|
||||
forbidden:
|
||||
production_deploy
|
||||
submit_owner_approval
|
||||
tag_publish
|
||||
baseline_registry_finalize
|
||||
runtime_switch
|
||||
|
||||
dependencies:
|
||||
credential_projection
|
||||
trusted_builder
|
||||
|
||||
health:
|
||||
authenticated_probe
|
||||
release_api_health
|
||||
worker_liveness
|
||||
```
|
||||
|
||||
Only a fully passing decision reaches:
|
||||
|
||||
```text
|
||||
EXECUTION_READY
|
||||
```
|
||||
|
||||
A failed decision remains:
|
||||
|
||||
```text
|
||||
EXECUTION_READY_FAILED
|
||||
```
|
||||
|
||||
and the worker does not call `/claim`.
|
||||
|
||||
Therefore an infrastructure defect discovered by admission does not consume a business attempt.
|
||||
|
||||
### System Invariants
|
||||
|
||||
The same admission checks are exposed as a machine-readable invariant snapshot.
|
||||
|
||||
Examples:
|
||||
|
||||
```text
|
||||
capability_contract_registry
|
||||
task_capabilities_registered
|
||||
worker_capabilities_registered
|
||||
capability_match
|
||||
forbidden_operations_absent
|
||||
credential_projection
|
||||
trusted_builder
|
||||
authenticated_probe
|
||||
release_api_health
|
||||
worker_liveness
|
||||
```
|
||||
|
||||
A runtime can expose:
|
||||
|
||||
```text
|
||||
HEALTHY
|
||||
DEGRADED
|
||||
```
|
||||
|
||||
without relying on a human reconstructing state from logs.
|
||||
|
||||
### Repair Policy
|
||||
|
||||
Failure classification maps to bounded recovery:
|
||||
|
||||
```text
|
||||
DEPENDENCY_UNAVAILABLE → level 0 / auto local
|
||||
RESOURCE_UNAVAILABLE → level 0 / auto local
|
||||
WORKER_CRASH → level 0 / auto local
|
||||
LEASE_LOST → level 0 / auto local
|
||||
|
||||
CAPABILITY_CONTRACT_ERROR → level 1 / same-lineage repair task
|
||||
CREDENTIAL_PROJECTION_ERROR → level 1 / same-lineage repair task
|
||||
UNKNOWN → level 1 / diagnostic repair task
|
||||
|
||||
POLICY_DENIED → level 2 / Owner authorization
|
||||
BUSINESS_PERMANENT → level 3 / terminal/manual decision
|
||||
```
|
||||
|
||||
Only business-permanent failure consumes the business retry budget in this first contract.
|
||||
|
||||
The current Task API does not yet have a first-class `EXECUTION_READY_FAILED` durable state. Until it
|
||||
does, the Release Worker enforces the most important invariant by refusing to claim a Task whose
|
||||
admission fails.
|
||||
|
||||
## Capability Contract Registry
|
||||
|
||||
`aios_capability_contract.py` remains backward-compatible with existing aliases but now validates
|
||||
every alias target against a versioned semantic contract registry.
|
||||
|
||||
Unknown capabilities and wildcard capabilities remain fail-closed.
|
||||
|
||||
Agent registration and Task creation should next be changed to call `registered_capability()` so a
|
||||
capability cannot be accepted at registration and rejected only later by the matcher.
|
||||
|
||||
## Credential rule
|
||||
|
||||
Admission probes credential usability, not merely credential existence.
|
||||
|
||||
For release candidate creation the authenticated probe intentionally sends an invalid empty
|
||||
`/createReleaseCandidate` request.
|
||||
|
||||
Expected:
|
||||
|
||||
```text
|
||||
HTTP 400
|
||||
```
|
||||
|
||||
This proves:
|
||||
|
||||
```text
|
||||
credential readable
|
||||
→ authentication succeeded
|
||||
→ scope accepted
|
||||
→ request validation rejected empty business input
|
||||
```
|
||||
|
||||
A permission failure/401/403/5xx is `EXECUTION_READY_FAILED`.
|
||||
|
||||
## Lineage
|
||||
|
||||
Repair actions are descriptive, not permission grants. Any repair Task must retain the existing
|
||||
root/parent/attempt/authority chain and use governed same-lineage recovery.
|
||||
|
||||
No repair action may silently create a replacement business Task.
|
||||
|
||||
## Authority separation
|
||||
|
||||
Capability answers:
|
||||
|
||||
```text
|
||||
can this runtime perform the operation?
|
||||
```
|
||||
|
||||
Authority answers:
|
||||
|
||||
```text
|
||||
is this operation currently authorized?
|
||||
```
|
||||
|
||||
`release:candidate` never implies Owner approval, deployment, tag publication, baseline finalization
|
||||
or runtime switch.
|
||||
|
||||
## Deployment sequence
|
||||
|
||||
Phase 1 (this candidate):
|
||||
|
||||
- Capability Contract Registry;
|
||||
- shared Autonomous Control Plane library;
|
||||
- Release Worker pre-claim admission;
|
||||
- failure classification + repair evidence.
|
||||
|
||||
Phase 2:
|
||||
|
||||
- Task API uses the same library before returning Tasks from poll and again during claim CAS;
|
||||
- Task schema gains durable `ADMITTED/EXECUTION_READY/EXECUTION_READY_FAILED` evidence without
|
||||
counting pre-side-effect infrastructure failures against business attempts.
|
||||
|
||||
Phase 3:
|
||||
|
||||
- Agent registration validates against the same Capability Contract Registry;
|
||||
- external connector writes only durable Intent/Command Inbox records and leaves execution to the
|
||||
local control plane.
|
||||
|
||||
This sequencing avoids a second controller stack while converging every decision onto one contract.
|
||||
@@ -0,0 +1,228 @@
|
||||
# Durable Owner Authorization and capability projection
|
||||
|
||||
## Problem
|
||||
|
||||
AIOS already supports unattended worker execution after a Task is executable, but Owner intent was
|
||||
not durable at the capability layer. The same Owner repeatedly had to authorize or repair:
|
||||
|
||||
- `governed-task:project`;
|
||||
- `release:candidate`.
|
||||
|
||||
This caused valid one-time Owner intent to be lost between conversations and forced repeated human
|
||||
interaction.
|
||||
|
||||
## Design
|
||||
|
||||
Durable authorization is represented by a root-owned mode `0600` policy outside Git. The policy is
|
||||
persistent until explicitly revoked.
|
||||
|
||||
The first persistable capabilities are deliberately narrow:
|
||||
|
||||
```text
|
||||
governed-task:project
|
||||
release:candidate
|
||||
```
|
||||
|
||||
They cover preparation work only.
|
||||
|
||||
The following capabilities can never be inherited from this durable grant:
|
||||
|
||||
```text
|
||||
submitOwnerApproval
|
||||
startDeployment
|
||||
tag_publish
|
||||
baseline_registry_finalize
|
||||
runtime_switch
|
||||
main_write
|
||||
production_deploy
|
||||
force_push
|
||||
credential_admin
|
||||
```
|
||||
|
||||
Those remain separate high-risk release/Owner-signature gates.
|
||||
|
||||
## Policy
|
||||
|
||||
Example schema:
|
||||
|
||||
```json
|
||||
{
|
||||
"schema": "aios.owner-durable-authorization.v1",
|
||||
"grant_id": "owner-permanent-preparation-20260819",
|
||||
"owner_identity": "owner@example",
|
||||
"account": "chatgpt-governed-implementer-singlehost",
|
||||
"capabilities": [
|
||||
"governed-task:project",
|
||||
"release:candidate"
|
||||
],
|
||||
"issued_at": "2026-08-19T05:00:00Z",
|
||||
"not_before": "2026-08-19T05:00:00Z",
|
||||
"persistent_until_revoked": true,
|
||||
"constraints": {
|
||||
"repository": "aios/aios",
|
||||
"allowed_target_prefix": "agent/",
|
||||
"release_version_prefix": "V2R2C",
|
||||
"max_task_risk": "engineering"
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
The repository never stores the live Owner policy or release credential.
|
||||
|
||||
## Revocation
|
||||
|
||||
A root-owned mode `0600` revocation file lists `revoked_grant_ids`.
|
||||
|
||||
Revocation is checked on every operation, not only at service startup. A running broker therefore
|
||||
fails closed immediately after revocation without needing a new conversation.
|
||||
|
||||
## Broker boundary
|
||||
|
||||
The broker exposes only an AF_UNIX socket:
|
||||
|
||||
```text
|
||||
/run/aios-owner-authorization-broker/broker.sock
|
||||
```
|
||||
|
||||
The socket is mode `0660` and owned by the service's `aios-executor-exchange` group. This avoids introducing another long-lived bearer secret solely for local broker access.
|
||||
|
||||
It exposes only:
|
||||
|
||||
```text
|
||||
GET /v1/capabilities
|
||||
POST /v1/project-governed-task
|
||||
POST /v1/authorize-release-candidate
|
||||
```
|
||||
|
||||
`project-governed-task` calls only the canonical Bridge.
|
||||
|
||||
`authorize-release-candidate` persists an Owner authorization decision only. It deliberately does
|
||||
**not** invent a local trusted-builder endpoint. The external release connector remains responsible
|
||||
for projecting the account's real `release:candidate` scope. Until that external projection exists,
|
||||
the decision state is `AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION`.
|
||||
|
||||
The broker holds only the existing host control credential needed for canonical governed-task
|
||||
projection. It never holds an Owner signing key or a release/deployment credential. The caller also supplies the expected account identifier over the Unix transport; the OS socket permission is the local access boundary.
|
||||
|
||||
|
||||
## Canonical Online Task projection
|
||||
|
||||
A durable `governed-task:project` grant may create a complete canonical Online Task rather than a
|
||||
bare Task. It requires:
|
||||
|
||||
```text
|
||||
require_online_task_projection=true
|
||||
generation=1
|
||||
branch_mode=CREATE_NEW
|
||||
max_attempts<=3
|
||||
risk<=engineering
|
||||
initial action=source_capture
|
||||
```
|
||||
|
||||
Projected task capabilities are limited to:
|
||||
|
||||
```text
|
||||
git.read
|
||||
git.test
|
||||
governed.write.branch
|
||||
governed.write.file
|
||||
```
|
||||
|
||||
A write-capable task still begins as `source_capture`. Existing placeholder-deferral prevents the
|
||||
generic worker from claiming it until a deterministic plan is registered. Repository mutation
|
||||
therefore still requires the normal deterministic plan, WriterLease and governed commit path.
|
||||
|
||||
This fixes the previous split-brain behavior where a bare Task could exist without the canonical
|
||||
governed execution projection.
|
||||
|
||||
|
||||
## External capability projection outbox
|
||||
|
||||
A `release:candidate` authorization now creates a durable machine-readable outbox record:
|
||||
|
||||
```text
|
||||
schema=aios.external-capability-projection.request.v1
|
||||
state=PENDING_EXTERNAL_CONNECTOR
|
||||
capability=release:candidate
|
||||
```
|
||||
|
||||
The projection identifier is deterministic from:
|
||||
|
||||
```text
|
||||
grant_id
|
||||
account
|
||||
capability
|
||||
request_sha256
|
||||
```
|
||||
|
||||
The outbox stores no credentials or Owner signing material. It retains only the already-authorized
|
||||
release-candidate request, policy/request hashes, account, grant id and timestamps.
|
||||
|
||||
The broker exposes read-only Unix-socket endpoints:
|
||||
|
||||
```text
|
||||
GET /v1/projections
|
||||
GET /v1/projections/{projection_id}
|
||||
```
|
||||
|
||||
These endpoints re-check the current Owner policy on every read. If the durable grant has been
|
||||
revoked, enumeration fails closed with `OWNER_GRANT_REVOKED`. The historical outbox record is
|
||||
retained for audit and is not silently deleted.
|
||||
|
||||
The repository intentionally does **not** provide a connector acknowledgement/write endpoint yet.
|
||||
No in-repository code can truthfully assert that the external ChatGPT connector account has gained
|
||||
`release:candidate`. Until the platform consumes the outbox and performs its own identity-scope
|
||||
projection, the only valid state is:
|
||||
|
||||
```text
|
||||
AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION
|
||||
PENDING_EXTERNAL_CONNECTOR
|
||||
```
|
||||
|
||||
A future external projector must preserve:
|
||||
|
||||
```text
|
||||
release:candidate
|
||||
!= submitOwnerApproval
|
||||
!= startDeployment
|
||||
!= tag_publish
|
||||
!= baseline_registry_finalize
|
||||
!= runtime_switch
|
||||
```
|
||||
|
||||
## Idempotency and audit
|
||||
|
||||
Every broker operation requires an idempotency key. A receipt SQLite database stores only:
|
||||
|
||||
- idempotency key;
|
||||
- capability;
|
||||
- request SHA-256;
|
||||
- response SHA-256;
|
||||
- sanitized upstream response;
|
||||
- timestamp.
|
||||
|
||||
An exact retry returns the same receipt. Reusing an idempotency key with different content fails.
|
||||
|
||||
## Unattended invariant
|
||||
|
||||
Once the Owner policy and narrow service credentials are provisioned:
|
||||
|
||||
```text
|
||||
Owner durable grant
|
||||
→ broker decision
|
||||
→ canonical governed task projection / release candidate preparation
|
||||
→ worker / trusted builder
|
||||
```
|
||||
|
||||
No repeated Owner interaction is required for those preparation capabilities.
|
||||
|
||||
Production execution remains:
|
||||
|
||||
```text
|
||||
release candidate
|
||||
→ immutable approval envelope
|
||||
→ Owner signature
|
||||
→ deployment
|
||||
```
|
||||
|
||||
The durable grant never substitutes for the Owner signature.
|
||||
@@ -0,0 +1,87 @@
|
||||
# Release candidate capability contract remediation
|
||||
|
||||
## Problem
|
||||
|
||||
The release-preparation worker can be correctly registered with the raw capability:
|
||||
|
||||
```text
|
||||
release:candidate
|
||||
```
|
||||
|
||||
while the canonical Task API matcher still rejects the worker.
|
||||
|
||||
Before this repair, the capability matrix is:
|
||||
|
||||
```text
|
||||
worker raw = ["release:candidate"]
|
||||
worker effective= []
|
||||
task required = ["exact:release:candidate"]
|
||||
eligible = false
|
||||
```
|
||||
|
||||
The Task therefore remains `QUEUED`, `attempt=0`, and cannot be claimed even though both the Task and
|
||||
worker use the same literal capability.
|
||||
|
||||
## Root cause
|
||||
|
||||
`aios_capability_contract.py` is deliberately fail-closed:
|
||||
|
||||
- task-side unknown names become `exact:<name>`;
|
||||
- worker-side unknown names are discarded;
|
||||
- wildcard worker capabilities are ignored.
|
||||
|
||||
`release:candidate` was absent from both canonical alias tables.
|
||||
|
||||
## Repair
|
||||
|
||||
Add exactly one semantic capability to both vocabularies:
|
||||
|
||||
```text
|
||||
release:candidate → release:candidate
|
||||
```
|
||||
|
||||
Do not add:
|
||||
|
||||
```text
|
||||
release:*
|
||||
submitOwnerApproval
|
||||
startDeployment
|
||||
tag_publish
|
||||
baseline_registry_finalize
|
||||
runtime_switch
|
||||
```
|
||||
|
||||
Unknown names and wildcards must continue to fail closed.
|
||||
|
||||
## Runtime recovery
|
||||
|
||||
After the governed remediation candidate is officially verified:
|
||||
|
||||
1. install the candidate capability-contract module as a higher-precedence Task API overlay;
|
||||
2. restart only `aios-agent-task-api.service`;
|
||||
3. require `/health` success;
|
||||
4. require the fixed existing release Task to remain the same Task/idempotency;
|
||||
5. let the already-active dedicated release worker poll and claim it automatically;
|
||||
6. require heartbeat, trusted-builder result and official verification;
|
||||
7. do not manually claim or create a replacement live Task.
|
||||
|
||||
The fixed V2R2C2 live Task may complete as the acceptance candidate it was originally bound to.
|
||||
Because this remediation is a later source commit, the first release candidate that includes the
|
||||
matcher repair should use the next immutable version (for example `V2R2C3`) rather than rewriting an
|
||||
already-created V2R2C2 candidate.
|
||||
|
||||
## Governance boundary
|
||||
|
||||
This repair affects only Task eligibility. It grants no credential and performs no release action by
|
||||
itself.
|
||||
|
||||
The dedicated release worker still requires its independent `release:candidate` credential, and the
|
||||
following remain separate:
|
||||
|
||||
```text
|
||||
Owner signed approval
|
||||
deployment
|
||||
tag publication
|
||||
baseline finalization
|
||||
runtime switch
|
||||
```
|
||||
@@ -0,0 +1,82 @@
|
||||
# Fixed V2R2C2 release lineage contract
|
||||
|
||||
## Why this contract exists
|
||||
|
||||
AIOS has multiple engineering commits in the same ancestry, and they serve different purposes.
|
||||
|
||||
The fixed V2R2C2 live Task was created by the dedicated Release Worker activation. Its immutable
|
||||
release source is therefore the Release Worker implementation candidate:
|
||||
|
||||
```text
|
||||
71e736b05fab13efc06af7a4fe2dc022c1cf9a85
|
||||
```
|
||||
|
||||
Later engineering commits are descendants but are not retroactive V2R2C2 release sources:
|
||||
|
||||
```text
|
||||
17deb91ca7a187dc81766945b1174e854f86bd25
|
||||
capability-contract remediation
|
||||
|
||||
b4b4416c93fedbe451d655d49e28f915b8fd38e6
|
||||
Autonomous Control Plane engineering candidate
|
||||
```
|
||||
|
||||
Conflating "current engineering base" with "immutable release source" caused ACP activation to
|
||||
fail closed with `FIXED_RELEASE_SOURCE_DRIFT`.
|
||||
|
||||
## Canonical fixed lineage
|
||||
|
||||
```text
|
||||
Task:
|
||||
r42-release-candidate-v2r2c2-live-20260819
|
||||
|
||||
candidate_id:
|
||||
r42-release-candidate-v2r2c2-20260819
|
||||
|
||||
source_commit:
|
||||
71e736b05fab13efc06af7a4fe2dc022c1cf9a85
|
||||
|
||||
version:
|
||||
V2R2C2
|
||||
|
||||
required_capabilities:
|
||||
["release:candidate"]
|
||||
```
|
||||
|
||||
This contract is immutable for V2R2C2.
|
||||
|
||||
## What may change
|
||||
|
||||
The following are lifecycle state, not release identity:
|
||||
|
||||
```text
|
||||
status
|
||||
attempt
|
||||
max_attempts
|
||||
revision
|
||||
claimed_by
|
||||
lease
|
||||
verification
|
||||
```
|
||||
|
||||
They may change through official retry/claim/result/verification without changing the release
|
||||
lineage.
|
||||
|
||||
## What must never be "repaired"
|
||||
|
||||
Do not rewrite the fixed V2R2C2 Task source to a later engineering commit.
|
||||
|
||||
A later engineering source that must itself be released receives a later immutable release version
|
||||
(for example V2R2C3), not a mutation of V2R2C2.
|
||||
|
||||
## Activation invariant
|
||||
|
||||
ACP activation must:
|
||||
|
||||
1. verify its own governed engineering candidate separately;
|
||||
2. read the live fixed V2R2C2 Task;
|
||||
3. validate that Task through `validate_fixed_v2r2c2_task()`;
|
||||
4. never compare the fixed release source to the ACP engineering base/candidate;
|
||||
5. never mutate the fixed Task merely to satisfy activation.
|
||||
|
||||
This makes engineering lineage and release lineage explicit and independently verifiable.
|
||||
@@ -0,0 +1,142 @@
|
||||
# Release Preparation Online Agent
|
||||
|
||||
## Purpose
|
||||
|
||||
This worker closes the release-preparation execution gap without widening the ChatGPT connector
|
||||
identity.
|
||||
|
||||
The dedicated Agent identity is:
|
||||
|
||||
```text
|
||||
online-release-worker-cn-network-mkt-20260819
|
||||
```
|
||||
|
||||
Its only advertised task capability is:
|
||||
|
||||
```text
|
||||
release:candidate
|
||||
```
|
||||
|
||||
The worker may create an immutable release candidate, but it never submits Owner approval and never
|
||||
starts deployment, publishes a tag, finalizes a baseline, switches runtime or writes Git.
|
||||
|
||||
## Credentials
|
||||
|
||||
The worker uses two independently scoped credentials:
|
||||
|
||||
```text
|
||||
agent_token
|
||||
release_candidate_token
|
||||
```
|
||||
|
||||
`agent_token` is the Online Agent registration credential. It must be encrypted into systemd's
|
||||
credential store and must never be written to Git, Task metadata, evidence or logs.
|
||||
|
||||
`release_candidate_token` is provisioned during Owner-authorized runtime activation. Only its SHA-256 record is stored in the release API token registry; the plaintext is immediately encrypted into systemd credstore and the temporary plaintext is removed. Its only allowed scope is:
|
||||
|
||||
```text
|
||||
release:candidate
|
||||
```
|
||||
|
||||
It is explicitly not an Owner signing key and must not grant:
|
||||
|
||||
```text
|
||||
submitOwnerApproval
|
||||
startDeployment
|
||||
tag_publish
|
||||
baseline_registry_finalize
|
||||
runtime_switch
|
||||
```
|
||||
|
||||
The unit reads both credentials through systemd `LoadCredentialEncrypted`; neither plaintext token is kept in the unit, repository, Task metadata, evidence, or logs.
|
||||
|
||||
## Task contract
|
||||
|
||||
The worker will only claim Tasks whose exact required capability set is:
|
||||
|
||||
```json
|
||||
["release:candidate"]
|
||||
```
|
||||
|
||||
The Task must provide a bounded `constraints.release_candidate` payload (metadata is accepted only for compatibility):
|
||||
|
||||
```json
|
||||
{
|
||||
"candidate_id": "...",
|
||||
"source_commit": "<40 hex>",
|
||||
"version": "V2R2C2",
|
||||
"projection_id": "projection-..."
|
||||
}
|
||||
```
|
||||
|
||||
No caller-supplied URL, method, headers, command, shell, environment, credential path, deployment
|
||||
operation or arbitrary release route is accepted.
|
||||
|
||||
The release API base is host configuration, not Task input. The only release mutation invoked by the
|
||||
worker is:
|
||||
|
||||
```text
|
||||
POST /createReleaseCandidate
|
||||
```
|
||||
|
||||
## Lifecycle
|
||||
|
||||
```text
|
||||
poll
|
||||
→ claim
|
||||
→ heartbeat
|
||||
→ createReleaseCandidate
|
||||
→ result submit
|
||||
→ official verifier
|
||||
```
|
||||
|
||||
The result declares:
|
||||
|
||||
```text
|
||||
writer_lease=NONE
|
||||
changed_files=[]
|
||||
direct_main_write=false
|
||||
production_write=false
|
||||
tag_write=false
|
||||
baseline_write=false
|
||||
runtime_switch=false
|
||||
owner_approval_submitted=false
|
||||
```
|
||||
|
||||
## Remaining high-risk gate
|
||||
|
||||
Successful release-candidate creation does not authorize Production.
|
||||
|
||||
The next stage remains:
|
||||
|
||||
```text
|
||||
release candidate
|
||||
→ immutable approval envelope
|
||||
→ Owner signature
|
||||
→ startDeployment
|
||||
```
|
||||
|
||||
That separation is mandatory even when release preparation is fully unattended.
|
||||
|
||||
|
||||
## Unattended availability
|
||||
|
||||
The release worker has its own systemd watchdog rather than modifying the shared worker-availability
|
||||
guard contract.
|
||||
|
||||
Every successful Task poll and Task heartbeat emits:
|
||||
|
||||
```text
|
||||
WATCHDOG=1
|
||||
```
|
||||
|
||||
and the service uses:
|
||||
|
||||
```text
|
||||
Type=notify
|
||||
WatchdogSec=90
|
||||
Restart=always
|
||||
```
|
||||
|
||||
A process that is technically alive but no longer polling/heartbeating therefore gets restarted by
|
||||
systemd. The worker also writes `worker-liveness.json` for operator diagnostics.
|
||||
@@ -0,0 +1 @@
|
||||
placeholder
|
||||
@@ -10,9 +10,13 @@ ROOT = Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0, str(ROOT / "active-runtime/app"))
|
||||
|
||||
from aios_capability_contract import ( # noqa: E402
|
||||
capability_contract,
|
||||
has_capabilities,
|
||||
normalize_required_capabilities,
|
||||
normalize_worker_capabilities,
|
||||
normalized_capability_matrix,
|
||||
registered_capability,
|
||||
validate_capability_contract_registry,
|
||||
)
|
||||
|
||||
|
||||
@@ -31,3 +35,32 @@ def test_runtime_capabilities_are_isolated_between_workers() -> None:
|
||||
|
||||
def test_unknown_runtime_capability_remains_fail_closed() -> None:
|
||||
assert has_capabilities(["agent.runtime.hermes"], ["agent.runtime.unknown"]) is False
|
||||
|
||||
|
||||
def test_autonomous_coordinator_capability_is_registered_on_both_sides() -> None:
|
||||
capability = "engineering.autonomous"
|
||||
assert registered_capability(capability) is True
|
||||
assert registered_capability(capability, worker=True) is True
|
||||
assert normalize_required_capabilities([capability]) == frozenset({capability})
|
||||
assert normalize_worker_capabilities([capability]) == frozenset({capability})
|
||||
contract = capability_contract(capability)
|
||||
assert contract is not None
|
||||
assert contract.dependencies == ("agent_task_control_plane",)
|
||||
assert validate_capability_contract_registry()["valid"] is True
|
||||
|
||||
|
||||
def test_supervisor_coordinator_requires_runtime_and_autonomous_capabilities() -> None:
|
||||
required = ["agent.runtime.openclaw", "engineering.autonomous"]
|
||||
assert normalized_capability_matrix(required, required) == {
|
||||
"worker_effective": required,
|
||||
"task_required": required,
|
||||
"eligible": True,
|
||||
}
|
||||
assert has_capabilities(["agent.runtime.openclaw"], required) is False
|
||||
assert has_capabilities(["engineering.autonomous"], required) is False
|
||||
|
||||
|
||||
def test_autonomous_capability_near_matches_remain_fail_closed() -> None:
|
||||
worker = ["agent.runtime.openclaw", "engineering.autonomous"]
|
||||
assert has_capabilities(worker, ["engineering.autonomous.extra"]) is False
|
||||
assert has_capabilities(worker, ["engineering.*"]) is False
|
||||
|
||||
@@ -0,0 +1,452 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from pathlib import Path
|
||||
import hashlib
|
||||
import json
|
||||
import sqlite3
|
||||
import sys
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0,str(ROOT/"active-runtime/app"))
|
||||
|
||||
from aios_agent_task_initial_execution_projection import (
|
||||
InitialExecutionProjectionError,
|
||||
InitialExecutionProjectionRequest,
|
||||
InitialExecutionProjectionStore,
|
||||
SourceCaptureProjectionRecoveryRequest,
|
||||
)
|
||||
from aios_online_agent_execution import (
|
||||
GovernedExecutionBinding,
|
||||
project_execution_envelope,
|
||||
)
|
||||
|
||||
AUTHORIZATION="owner-initial-execution-projection-recovery-20260820"
|
||||
TASK_ID="task-initial-projection-a"
|
||||
JOB_ID="job-initial-projection-a"
|
||||
BASE="a"*40
|
||||
BRANCH="agent/task-initial-projection-a"
|
||||
IDEMPOTENCY="task-initial-projection-binding-v1"
|
||||
CAPABILITY="agent.runtime.openclaw"
|
||||
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def _auth(monkeypatch):
|
||||
monkeypatch.setenv(
|
||||
"AIOS_AGENT_TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZATION_ID",
|
||||
AUTHORIZATION,
|
||||
)
|
||||
|
||||
|
||||
def canon(value):
|
||||
return hashlib.sha256(
|
||||
json.dumps(value,sort_keys=True,separators=(",",":")).encode()
|
||||
).hexdigest()
|
||||
|
||||
|
||||
def binding():
|
||||
return GovernedExecutionBinding.from_mapping({
|
||||
"task_id":TASK_ID,
|
||||
"job_id":JOB_ID,
|
||||
"repository":"aios/aios",
|
||||
"base_oid":BASE,
|
||||
"expected_branch_oid":BASE,
|
||||
"target_branch":BRANCH,
|
||||
"branch_mode":"CREATE_NEW",
|
||||
"generation":1,
|
||||
"idempotency_key":IDEMPOTENCY,
|
||||
})
|
||||
|
||||
|
||||
def runtime_input():
|
||||
return {
|
||||
"goal":"Supervise release preparation",
|
||||
"instructions":"Observe and resume only the same lineage.",
|
||||
"success_criteria":[{"id":"heartbeat","description":"Heartbeat observed"}],
|
||||
"constraints":{"production_write":False},
|
||||
"scope":{"scope_type":"legacy","tenant_id":None,"project_id":None},
|
||||
"trace_id":TASK_ID,
|
||||
}
|
||||
|
||||
|
||||
def envelope(*, job_id=JOB_ID, action="agent_runtime", capability=CAPABILITY):
|
||||
bound=binding()
|
||||
if job_id != JOB_ID:
|
||||
bound=GovernedExecutionBinding.from_mapping({**bound.public(),"job_id":job_id})
|
||||
return project_execution_envelope(
|
||||
bound,
|
||||
{
|
||||
"action":action,
|
||||
"adapter":"openclaw" if action=="agent_runtime" else "source_capture",
|
||||
"capability_scope":[capability],
|
||||
"runtime_input":runtime_input() if action=="agent_runtime" else None,
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
def make_db(tmp_path, *, claimed_by=None, lease=False, metadata=None, caps=None, event="TASK_CREATED"):
|
||||
path=tmp_path/"state.sqlite3"
|
||||
con=sqlite3.connect(path)
|
||||
con.executescript("""
|
||||
CREATE TABLE tasks(
|
||||
task_id TEXT PRIMARY KEY,
|
||||
goal TEXT NOT NULL,
|
||||
instructions TEXT NOT NULL,
|
||||
success_criteria_json TEXT NOT NULL,
|
||||
required_capabilities_json TEXT NOT NULL,
|
||||
constraints_json TEXT NOT NULL,
|
||||
metadata_json TEXT NOT NULL,
|
||||
deadline TEXT,
|
||||
priority INTEGER NOT NULL,
|
||||
max_attempts INTEGER NOT NULL,
|
||||
attempt INTEGER NOT NULL DEFAULT 0,
|
||||
status TEXT NOT NULL,
|
||||
claimed_by TEXT,
|
||||
lease_token_hash TEXT,
|
||||
lease_expires_at TEXT,
|
||||
result_json TEXT,
|
||||
result_digest TEXT,
|
||||
result_idempotency_key TEXT,
|
||||
verification_json TEXT,
|
||||
revision INTEGER NOT NULL DEFAULT 1,
|
||||
created_at TEXT NOT NULL,
|
||||
updated_at TEXT NOT NULL
|
||||
);
|
||||
CREATE TABLE events(
|
||||
event_id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
task_id TEXT,
|
||||
agent_id TEXT,
|
||||
event_type TEXT NOT NULL,
|
||||
payload_json TEXT NOT NULL,
|
||||
created_at TEXT NOT NULL
|
||||
);
|
||||
""")
|
||||
metadata={} if metadata is None else metadata
|
||||
caps=[] if caps is None else caps
|
||||
con.execute(
|
||||
"INSERT INTO tasks VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)",
|
||||
(
|
||||
TASK_ID,"Supervise release preparation",
|
||||
"Observe and resume only the same lineage.",
|
||||
json.dumps([{"id":"heartbeat","description":"Heartbeat observed"}]),
|
||||
json.dumps(caps),json.dumps({"production_write":False}),json.dumps(metadata),
|
||||
None,100,20,0,"QUEUED",claimed_by,
|
||||
"lease-hash" if lease else None,
|
||||
"2099-01-01T00:00:00Z" if lease else None,
|
||||
None,None,None,None,1,"2026-08-20T00:00:00Z","2026-08-20T00:00:00Z",
|
||||
),
|
||||
)
|
||||
con.execute(
|
||||
"INSERT INTO events(task_id,agent_id,event_type,payload_json,created_at) VALUES (?,?,?,?,?)",
|
||||
(TASK_ID,None,event,"{}","2026-08-20T00:00:00Z"),
|
||||
)
|
||||
con.commit();con.close()
|
||||
return path
|
||||
|
||||
|
||||
def request(*, digest=None, key="initial-projection-a-v1"):
|
||||
return InitialExecutionProjectionRequest.from_mapping({
|
||||
"expected_status":"QUEUED",
|
||||
"expected_revision":1,
|
||||
"expected_attempt":0,
|
||||
"expected_metadata_sha256":canon({}),
|
||||
"expected_required_capabilities_sha256":canon([]),
|
||||
"expected_job_id":JOB_ID,
|
||||
"expected_execution_digest":digest or envelope()["execution_digest"],
|
||||
"required_capability":CAPABILITY,
|
||||
"adapter":"openclaw",
|
||||
"authorization_id":AUTHORIZATION,
|
||||
"idempotency_key":key,
|
||||
"reason":"initial governed supervisor runtime projection",
|
||||
})
|
||||
|
||||
|
||||
def resolver(_task_id,_job_id,_capability,_adapter,received_runtime_input):
|
||||
assert received_runtime_input==runtime_input()
|
||||
return envelope()
|
||||
|
||||
|
||||
def source_capture_result():
|
||||
old=envelope(action="source_capture",capability="git.read")
|
||||
return {
|
||||
"task_id":TASK_ID,
|
||||
"job_id":JOB_ID,
|
||||
"repository":"aios/aios",
|
||||
"base_oid":BASE,
|
||||
"expected_branch_oid":BASE,
|
||||
"target_branch":BRANCH,
|
||||
"generation":1,
|
||||
"status":"SUCCEEDED",
|
||||
"classification":"NEEDS_REWORK_SOURCE_CAPTURE",
|
||||
"execution_mode":"source_capture",
|
||||
"execution_digest":old["execution_digest"],
|
||||
"changed_files":[],
|
||||
"writer_lease":"NONE",
|
||||
"verification_state":"NEEDS_REWORK",
|
||||
"direct_main_write":False,
|
||||
"production_write":False,
|
||||
"force_push":False,
|
||||
"protected_ref_bypass":False,
|
||||
"secret_material_exposed":False,
|
||||
"source_snapshot":{"schema":"test.snapshot.v1","payload_sha256":"d"*64},
|
||||
}
|
||||
|
||||
|
||||
def make_source_capture_rework_db(tmp_path, *, mutate=None, events=None):
|
||||
path=make_db(tmp_path)
|
||||
result=source_capture_result()
|
||||
verification={
|
||||
"decision":"NEEDS_REWORK",
|
||||
"reason":"SOURCE_CAPTURE_COMPLETE_RUNTIME_PROJECTION_REQUIRED",
|
||||
}
|
||||
if mutate:
|
||||
mutate(result,verification)
|
||||
event_types=events or [
|
||||
"TASK_CLAIMED","TASK_HEARTBEAT","RESULT_SUBMITTED","TASK_VERIFIED"
|
||||
]
|
||||
con=sqlite3.connect(path)
|
||||
con.execute(
|
||||
"UPDATE tasks SET attempt=1,status='QUEUED',claimed_by=NULL,revision=5,"
|
||||
"result_json=?,result_digest=?,result_idempotency_key=?,verification_json=?",
|
||||
(
|
||||
json.dumps(result,sort_keys=True,separators=(",",":")),
|
||||
canon(result),"source-capture-result-v1",
|
||||
json.dumps(verification,sort_keys=True,separators=(",",":")),
|
||||
),
|
||||
)
|
||||
for event_type in event_types:
|
||||
payload={}
|
||||
if event_type=="TASK_VERIFIED":
|
||||
payload={"decision":"NEEDS_REWORK","next_state":"QUEUED"}
|
||||
con.execute(
|
||||
"INSERT INTO events(task_id,agent_id,event_type,payload_json,created_at) "
|
||||
"VALUES (?,?,?,?,?)",
|
||||
(TASK_ID,"generic-worker",event_type,json.dumps(payload),"2026-08-20T00:01:00Z"),
|
||||
)
|
||||
con.commit();con.close()
|
||||
return path,result,verification
|
||||
|
||||
|
||||
def recovery_request(result,verification,*,digest=None,key="source-capture-recovery-a-v1"):
|
||||
return SourceCaptureProjectionRecoveryRequest.from_mapping({
|
||||
"expected_status":"QUEUED",
|
||||
"expected_revision":5,
|
||||
"expected_attempt":1,
|
||||
"expected_metadata_sha256":canon({}),
|
||||
"expected_required_capabilities_sha256":canon([]),
|
||||
"previous_verification_sha256":canon(verification),
|
||||
"expected_source_capture_result_sha256":canon(result),
|
||||
"expected_source_capture_execution_digest":result["execution_digest"],
|
||||
"expected_job_id":JOB_ID,
|
||||
"expected_execution_digest":digest or envelope()["execution_digest"],
|
||||
"required_capability":CAPABILITY,
|
||||
"adapter":"openclaw",
|
||||
"authorization_id":AUTHORIZATION,
|
||||
"idempotency_key":key,
|
||||
"reason":"first durable runtime projection after governed source capture",
|
||||
})
|
||||
|
||||
|
||||
def test_initial_projection_cas_persists_envelope_capability_and_event(tmp_path):
|
||||
path=make_db(tmp_path)
|
||||
result=InitialExecutionProjectionStore(path,resolver).project(TASK_ID,request())
|
||||
assert result["revision"]==2
|
||||
assert result["attempt"]==0
|
||||
assert result["idempotent"] is False
|
||||
con=sqlite3.connect(path)
|
||||
row=con.execute(
|
||||
"SELECT status,attempt,revision,claimed_by,metadata_json,required_capabilities_json FROM tasks"
|
||||
).fetchone()
|
||||
assert row[:4]==("QUEUED",0,2,None)
|
||||
metadata=json.loads(row[4])
|
||||
assert metadata["execution"]["action"]=="agent_runtime"
|
||||
assert metadata["execution"]["adapter"]=="openclaw"
|
||||
assert metadata["execution"]["governance"]["writer_lease"] is False
|
||||
assert json.loads(row[5])==[CAPABILITY]
|
||||
events=con.execute("SELECT event_type,payload_json FROM events ORDER BY event_id").fetchall()
|
||||
assert [event[0] for event in events]==[
|
||||
"TASK_CREATED","TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZED"
|
||||
]
|
||||
payload=json.loads(events[-1][1])
|
||||
assert payload["job_id"]==JOB_ID
|
||||
assert payload["main_write"] is False
|
||||
assert payload["production_write"] is False
|
||||
|
||||
|
||||
def test_initial_projection_idempotent_exact_replay(tmp_path):
|
||||
path=make_db(tmp_path)
|
||||
store=InitialExecutionProjectionStore(path,resolver)
|
||||
first=store.project(TASK_ID,request())
|
||||
second=store.project(TASK_ID,request())
|
||||
assert first["idempotent"] is False
|
||||
assert second["idempotent"] is True
|
||||
assert second["execution_digest"]==first["execution_digest"]
|
||||
con=sqlite3.connect(path)
|
||||
assert con.execute("SELECT COUNT(*) FROM task_initial_execution_projections").fetchone()[0]==1
|
||||
assert con.execute("SELECT COUNT(*) FROM events").fetchone()[0]==2
|
||||
|
||||
|
||||
def test_source_capture_recovery_projects_first_runtime_envelope(tmp_path):
|
||||
path,result,verification=make_source_capture_rework_db(tmp_path)
|
||||
response=InitialExecutionProjectionStore(path,resolver).project_from_source_capture(
|
||||
TASK_ID,recovery_request(result,verification)
|
||||
)
|
||||
assert response["revision"]==6
|
||||
assert response["attempt"]==1
|
||||
assert response["recovery_from_source_capture"] is True
|
||||
assert response["idempotent"] is False
|
||||
con=sqlite3.connect(path)
|
||||
row=con.execute(
|
||||
"SELECT status,attempt,revision,claimed_by,metadata_json,"
|
||||
"required_capabilities_json,result_json,verification_json FROM tasks"
|
||||
).fetchone()
|
||||
assert row[:4]==("QUEUED",1,6,None)
|
||||
metadata=json.loads(row[4])
|
||||
assert metadata["execution"]["action"]=="agent_runtime"
|
||||
assert metadata["execution"]["adapter"]=="openclaw"
|
||||
assert metadata["execution"]["governance"]["writer_lease"] is False
|
||||
assert json.loads(row[5])==[CAPABILITY]
|
||||
assert json.loads(row[6])==result
|
||||
assert json.loads(row[7])==verification
|
||||
events=con.execute("SELECT event_type,payload_json FROM events ORDER BY event_id").fetchall()
|
||||
assert [event[0] for event in events]==[
|
||||
"TASK_CREATED","TASK_CLAIMED","TASK_HEARTBEAT","RESULT_SUBMITTED",
|
||||
"TASK_VERIFIED","TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZED",
|
||||
]
|
||||
payload=json.loads(events[-1][1])
|
||||
assert payload["recovery_from_source_capture"] is True
|
||||
assert payload["source_capture_execution_digest"]==result["execution_digest"]
|
||||
assert payload["main_write"] is False
|
||||
assert payload["production_write"] is False
|
||||
|
||||
|
||||
def test_source_capture_recovery_idempotent_exact_replay(tmp_path):
|
||||
path,result,verification=make_source_capture_rework_db(tmp_path)
|
||||
store=InitialExecutionProjectionStore(path,resolver)
|
||||
first=store.project_from_source_capture(TASK_ID,recovery_request(result,verification))
|
||||
second=store.project_from_source_capture(TASK_ID,recovery_request(result,verification))
|
||||
assert first["idempotent"] is False
|
||||
assert second["idempotent"] is True
|
||||
assert second["execution_digest"]==first["execution_digest"]
|
||||
con=sqlite3.connect(path)
|
||||
assert con.execute(
|
||||
"SELECT COUNT(*) FROM task_source_capture_projection_recoveries"
|
||||
).fetchone()[0]==1
|
||||
assert con.execute("SELECT COUNT(*) FROM events").fetchone()[0]==6
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"mutate,error",
|
||||
[
|
||||
(lambda result,verification: result.__setitem__("changed_files",["x"]),
|
||||
"SOURCE_CAPTURE_RESULT_CONTRACT_INVALID"),
|
||||
(lambda result,verification: result.__setitem__("job_id","job-other"),
|
||||
"SOURCE_CAPTURE_RESULT_CONTRACT_INVALID"),
|
||||
(lambda result,verification: verification.__setitem__("decision","SUCCEEDED"),
|
||||
"SOURCE_CAPTURE_NEEDS_REWORK_VERIFICATION_REQUIRED"),
|
||||
],
|
||||
)
|
||||
def test_source_capture_recovery_rejects_unsafe_result_or_verification(
|
||||
tmp_path,mutate,error
|
||||
):
|
||||
path,result,verification=make_source_capture_rework_db(tmp_path,mutate=mutate)
|
||||
with pytest.raises(InitialExecutionProjectionError,match=error):
|
||||
InitialExecutionProjectionStore(path,resolver).project_from_source_capture(
|
||||
TASK_ID,recovery_request(result,verification)
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"events",
|
||||
[
|
||||
["TASK_CLAIMED","RESULT_SUBMITTED","TASK_VERIFIED"],
|
||||
["TASK_CLAIMED","TASK_HEARTBEAT","RESULT_SUBMITTED","TASK_VERIFIED","TASK_CLAIMED"],
|
||||
["TASK_CLAIMED","TASK_HEARTBEAT","TASK_VERIFIED","RESULT_SUBMITTED"],
|
||||
],
|
||||
)
|
||||
def test_source_capture_recovery_rejects_invalid_event_chain(tmp_path,events):
|
||||
path,result,verification=make_source_capture_rework_db(tmp_path,events=events)
|
||||
with pytest.raises(
|
||||
InitialExecutionProjectionError,match="SOURCE_CAPTURE_RECOVERY_EVENT_CHAIN_INVALID"
|
||||
):
|
||||
InitialExecutionProjectionStore(path,resolver).project_from_source_capture(
|
||||
TASK_ID,recovery_request(result,verification)
|
||||
)
|
||||
|
||||
|
||||
def test_source_capture_recovery_rejects_execution_digest_drift(tmp_path):
|
||||
path,result,verification=make_source_capture_rework_db(tmp_path)
|
||||
with pytest.raises(InitialExecutionProjectionError,match="EXECUTION_DIGEST_DRIFT"):
|
||||
InitialExecutionProjectionStore(path,resolver).project_from_source_capture(
|
||||
TASK_ID,recovery_request(result,verification,digest="f"*64)
|
||||
)
|
||||
|
||||
|
||||
def test_source_capture_recovery_request_is_strict(tmp_path):
|
||||
_path,result,verification=make_source_capture_rework_db(tmp_path)
|
||||
data=recovery_request(result,verification).public()
|
||||
data["metadata_patch"]={"execution":"forbidden"}
|
||||
with pytest.raises(InitialExecutionProjectionError,match="FIELDS_INVALID"):
|
||||
SourceCaptureProjectionRecoveryRequest.from_mapping(data)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"kwargs,error",
|
||||
[
|
||||
({"claimed_by":"worker-a"},"TASK_ALREADY_CLAIMED"),
|
||||
({"lease":True},"ACTIVE_LEASE_DENY"),
|
||||
({"metadata":{"execution":{}}},"TASK_EXECUTION_ALREADY_PROJECTED"),
|
||||
({"caps":["git.read"]},"TASK_INITIAL_CAPABILITIES_MUST_BE_EMPTY"),
|
||||
({"event":"TASK_CLAIMED"},"TASK_INITIAL_EVENT_CHAIN_REQUIRED"),
|
||||
],
|
||||
)
|
||||
def test_initial_projection_rejects_noninitial_state(tmp_path,kwargs,error):
|
||||
path=make_db(tmp_path,**kwargs)
|
||||
with pytest.raises(InitialExecutionProjectionError,match=error):
|
||||
InitialExecutionProjectionStore(path,resolver).project(TASK_ID,request())
|
||||
|
||||
|
||||
def test_initial_projection_rejects_binding_drift(tmp_path):
|
||||
path=make_db(tmp_path)
|
||||
def wrong(*_args):
|
||||
return envelope(job_id="job-other")
|
||||
with pytest.raises(InitialExecutionProjectionError,match="GOVERNED_BINDING_CONFLICT"):
|
||||
InitialExecutionProjectionStore(path,wrong).project(TASK_ID,request())
|
||||
|
||||
|
||||
def test_initial_projection_rejects_nonruntime_envelope(tmp_path):
|
||||
path=make_db(tmp_path)
|
||||
def wrong(*_args):
|
||||
return envelope(action="source_capture",capability="git.read")
|
||||
with pytest.raises(InitialExecutionProjectionError):
|
||||
InitialExecutionProjectionStore(path,wrong).project(TASK_ID,request())
|
||||
|
||||
|
||||
def test_initial_projection_rejects_execution_digest_drift(tmp_path):
|
||||
path=make_db(tmp_path)
|
||||
with pytest.raises(InitialExecutionProjectionError,match="EXECUTION_DIGEST_DRIFT"):
|
||||
InitialExecutionProjectionStore(path,resolver).project(
|
||||
TASK_ID,request(digest="f"*64)
|
||||
)
|
||||
|
||||
|
||||
def test_request_is_narrow_and_initial_state_bound():
|
||||
data=request().public()
|
||||
data["arbitrary_metadata"]={"x":1}
|
||||
with pytest.raises(InitialExecutionProjectionError,match="FIELDS_INVALID"):
|
||||
InitialExecutionProjectionRequest.from_mapping(data)
|
||||
data=request().public();data["expected_attempt"]=1
|
||||
with pytest.raises(InitialExecutionProjectionError,match="INITIAL_STATE_REQUIRED"):
|
||||
InitialExecutionProjectionRequest.from_mapping(data)
|
||||
|
||||
|
||||
def test_initial_projection_unit_is_unprivileged_and_owner_credential_bound():
|
||||
unit=(ROOT/"deploy/systemd/aios-agent-task-initial-execution-projection.service").read_text()
|
||||
assert "User=aios-agent-task" in unit
|
||||
assert "Group=aios-agent-task" in unit
|
||||
assert "LoadCredential=agent_task_control_token:" in unit
|
||||
assert "Environment=AIOS_GOVERNED_WRITE_API_URL=http://127.0.0.1:18892" in unit
|
||||
assert "RestrictAddressFamilies=AF_UNIX AF_INET" in unit
|
||||
assert "ReadWritePaths=/var/lib/aios-agent-task" in unit
|
||||
assert "User=root" not in unit
|
||||
@@ -0,0 +1,188 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from pathlib import Path
|
||||
import hashlib
|
||||
import json
|
||||
import sqlite3
|
||||
import sys
|
||||
|
||||
import pytest
|
||||
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0,str(ROOT/"active-runtime/app"))
|
||||
|
||||
from aios_agent_task_initial_execution_projection import InitialExecutionProjectionError
|
||||
from aios_agent_task_supervisor_coordinator_projection import (
|
||||
COORDINATOR_CAPABILITIES,
|
||||
RUNTIME_CAPABILITY,
|
||||
SupervisorCoordinatorProjectionRequest,
|
||||
SupervisorCoordinatorProjectionStore,
|
||||
)
|
||||
from aios_online_agent_execution import (
|
||||
GovernedExecutionBinding,
|
||||
project_execution_envelope,
|
||||
)
|
||||
|
||||
|
||||
TASK="task-supervisor-coordinator-a"
|
||||
JOB="job-supervisor-coordinator-a"
|
||||
BASE="a"*40
|
||||
BRANCH="agent/task-supervisor-coordinator-a"
|
||||
BINDING_KEY="task-supervisor-coordinator-binding-v1"
|
||||
AUTH="owner-supervisor-coordinator-projection-recovery-20260820"
|
||||
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def authorization(monkeypatch):
|
||||
monkeypatch.setenv(
|
||||
"AIOS_AGENT_TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZATION_ID",AUTH
|
||||
)
|
||||
|
||||
|
||||
def sha(value):
|
||||
return hashlib.sha256(
|
||||
json.dumps(value,sort_keys=True,separators=(",",":")).encode()
|
||||
).hexdigest()
|
||||
|
||||
|
||||
def binding():
|
||||
return GovernedExecutionBinding.from_mapping({
|
||||
"task_id":TASK,"job_id":JOB,"repository":"aios/aios",
|
||||
"base_oid":BASE,"expected_branch_oid":BASE,"target_branch":BRANCH,
|
||||
"branch_mode":"CREATE_NEW","generation":1,"idempotency_key":BINDING_KEY,
|
||||
})
|
||||
|
||||
|
||||
def runtime_input():
|
||||
return {
|
||||
"goal":"Durably supervise release preparation",
|
||||
"instructions":"Wait for exact connector ACK and preserve the same lineage.",
|
||||
"success_criteria":[{"id":"heartbeat","description":"Coordinator heartbeat"}],
|
||||
"constraints":{},
|
||||
"scope":{"scope_type":"legacy","tenant_id":None,"project_id":None},
|
||||
"trace_id":TASK,
|
||||
}
|
||||
|
||||
|
||||
def envelope(capabilities):
|
||||
return project_execution_envelope(binding(),{
|
||||
"action":"agent_runtime","adapter":"openclaw",
|
||||
"capability_scope":capabilities,"runtime_input":runtime_input(),
|
||||
})
|
||||
|
||||
|
||||
def planner_result(old_digest,*,access=False):
|
||||
return {
|
||||
"task_id":TASK,"job_id":JOB,"repository":"aios/aios",
|
||||
"base_oid":BASE,"expected_branch_oid":BASE,"target_branch":BRANCH,
|
||||
"generation":1,"status":"SUCCEEDED","execution_mode":"agent_runtime",
|
||||
"agent_runtime":"openclaw","execution_digest":old_digest,
|
||||
"control_plane_access_granted_to_runtime":access,
|
||||
"credential_material_exposed":False,"secret_material_exposed":False,
|
||||
"changed_files":[],"writer_lease":"NONE","direct_main_write":False,
|
||||
"production_write":False,"force_push":False,"protected_ref_bypass":False,
|
||||
"classification":"EXTERNAL_AGENT_RUNTIME_SUCCEEDED",
|
||||
}
|
||||
|
||||
|
||||
def make_db(tmp_path,*,access=False,heartbeat=True):
|
||||
tmp_path.mkdir(parents=True,exist_ok=True)
|
||||
path=tmp_path/"state.sqlite3";con=sqlite3.connect(path)
|
||||
con.executescript("""
|
||||
CREATE TABLE tasks(
|
||||
task_id TEXT PRIMARY KEY,goal TEXT NOT NULL,instructions TEXT NOT NULL,
|
||||
success_criteria_json TEXT NOT NULL,required_capabilities_json TEXT NOT NULL,
|
||||
constraints_json TEXT NOT NULL,metadata_json TEXT NOT NULL,deadline TEXT,
|
||||
priority INTEGER NOT NULL,max_attempts INTEGER NOT NULL,attempt INTEGER NOT NULL,
|
||||
status TEXT NOT NULL,claimed_by TEXT,lease_token_hash TEXT,lease_expires_at TEXT,
|
||||
result_json TEXT,result_digest TEXT,result_idempotency_key TEXT,
|
||||
verification_json TEXT,revision INTEGER NOT NULL,created_at TEXT,updated_at TEXT
|
||||
);
|
||||
CREATE TABLE events(
|
||||
event_id INTEGER PRIMARY KEY AUTOINCREMENT,task_id TEXT,agent_id TEXT,
|
||||
event_type TEXT NOT NULL,payload_json TEXT NOT NULL,created_at TEXT NOT NULL
|
||||
);
|
||||
""")
|
||||
old=envelope([RUNTIME_CAPABILITY]);result=planner_result(old["execution_digest"],access=access)
|
||||
verification={"decision":"NEEDS_REWORK","reason":"HOST_COORDINATOR_REQUIRED"}
|
||||
metadata={"execution":old,"execution_projection":{
|
||||
"schema":"aios.online-agent.task-projection.v1","task_id":TASK,
|
||||
"job_id":JOB,"execution_digest":old["execution_digest"],
|
||||
}}
|
||||
con.execute("INSERT INTO tasks VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)",(
|
||||
TASK,runtime_input()["goal"],runtime_input()["instructions"],
|
||||
json.dumps(runtime_input()["success_criteria"]),json.dumps([RUNTIME_CAPABILITY]),
|
||||
"{}",json.dumps(metadata,sort_keys=True,separators=(",",":")),None,100,20,2,
|
||||
"QUEUED",None,None,None,json.dumps(result,sort_keys=True,separators=(",",":")),
|
||||
sha(result),"planner-result-v1",json.dumps(verification,sort_keys=True,separators=(",",":")),
|
||||
11,"2026-08-20T00:00:00Z","2026-08-20T00:00:00Z",
|
||||
))
|
||||
events=["TASK_CREATED","TASK_INITIAL_EXECUTION_PROJECTION_AUTHORIZED","TASK_CLAIMED"]
|
||||
if heartbeat: events.append("TASK_HEARTBEAT")
|
||||
events += ["RESULT_SUBMITTED","TASK_VERIFIED"]
|
||||
for event in events:
|
||||
payload={"decision":"NEEDS_REWORK","next_state":"QUEUED"} if event=="TASK_VERIFIED" else {}
|
||||
con.execute("INSERT INTO events(task_id,agent_id,event_type,payload_json,created_at) VALUES (?,?,?,?,?)",(TASK,"worker",event,json.dumps(payload),"2026-08-20T00:00:00Z"))
|
||||
con.commit();con.close()
|
||||
return path,metadata,result,verification,old
|
||||
|
||||
|
||||
def request(metadata,result,verification,old,new):
|
||||
return SupervisorCoordinatorProjectionRequest.from_mapping({
|
||||
"expected_status":"QUEUED","expected_revision":11,"expected_attempt":2,
|
||||
"expected_metadata_sha256":sha(metadata),
|
||||
"expected_required_capabilities_sha256":sha([RUNTIME_CAPABILITY]),
|
||||
"previous_verification_sha256":sha(verification),
|
||||
"expected_planner_result_sha256":sha(result),
|
||||
"expected_old_execution_digest":old["execution_digest"],
|
||||
"expected_new_execution_digest":new["execution_digest"],
|
||||
"expected_job_id":JOB,"authorization_id":AUTH,
|
||||
"idempotency_key":"task-supervisor-coordinator-projection-v1",
|
||||
"reason":"credential-isolated host coordinator required",
|
||||
})
|
||||
|
||||
|
||||
def test_projects_exact_coordinator_capabilities_and_preserves_planner_evidence(tmp_path):
|
||||
path,metadata,result,verification,old=make_db(tmp_path)
|
||||
new=envelope(COORDINATOR_CAPABILITIES)
|
||||
store=SupervisorCoordinatorProjectionStore(path,lambda task,job,runtime:new)
|
||||
response=store.project(TASK,request(metadata,result,verification,old,new))
|
||||
assert response["revision"]==12 and response["attempt"]==2
|
||||
assert response["required_capabilities"]==COORDINATOR_CAPABILITIES
|
||||
assert response["credential_forwarded_to_planner"] is False
|
||||
con=sqlite3.connect(path)
|
||||
row=con.execute("select status,revision,metadata_json,required_capabilities_json,result_json from tasks").fetchone()
|
||||
assert row[:2]==("QUEUED",12)
|
||||
assert json.loads(row[2])["execution"]["execution_digest"]==new["execution_digest"]
|
||||
assert json.loads(row[3])==COORDINATOR_CAPABILITIES
|
||||
assert json.loads(row[4])==result
|
||||
assert con.execute("select event_type from events order by event_id desc limit 1").fetchone()[0]=="TASK_SUPERVISOR_COORDINATOR_PROJECTION_AUTHORIZED"
|
||||
|
||||
|
||||
def test_exact_replay_is_idempotent(tmp_path):
|
||||
path,metadata,result,verification,old=make_db(tmp_path);new=envelope(COORDINATOR_CAPABILITIES)
|
||||
store=SupervisorCoordinatorProjectionStore(path,lambda task,job,runtime:new)
|
||||
req=request(metadata,result,verification,old,new)
|
||||
assert store.project(TASK,req)["idempotent"] is False
|
||||
assert store.project(TASK,req)["idempotent"] is True
|
||||
|
||||
|
||||
def test_rejects_planner_with_control_access_or_missing_heartbeat(tmp_path):
|
||||
path,metadata,result,verification,old=make_db(tmp_path,access=True);new=envelope(COORDINATOR_CAPABILITIES)
|
||||
with pytest.raises(InitialExecutionProjectionError,match="RESULT_CONTRACT_INVALID"):
|
||||
SupervisorCoordinatorProjectionStore(path,lambda *args:new).project(
|
||||
TASK,request(metadata,result,verification,old,new)
|
||||
)
|
||||
path,metadata,result,verification,old=make_db(tmp_path/"second",heartbeat=False)
|
||||
with pytest.raises(InitialExecutionProjectionError,match="EVENT_CHAIN_INVALID"):
|
||||
SupervisorCoordinatorProjectionStore(path,lambda *args:new).project(
|
||||
TASK,request(metadata,result,verification,old,new)
|
||||
)
|
||||
|
||||
|
||||
def test_request_rejects_arbitrary_metadata_patch(tmp_path):
|
||||
path,metadata,result,verification,old=make_db(tmp_path);new=envelope(COORDINATOR_CAPABILITIES)
|
||||
value=request(metadata,result,verification,old,new).public();value["metadata_patch"]={}
|
||||
with pytest.raises(InitialExecutionProjectionError,match="FIELDS_INVALID"):
|
||||
SupervisorCoordinatorProjectionRequest.from_mapping(value)
|
||||
@@ -0,0 +1,157 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import sys
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
APP=ROOT/"active-runtime/app"
|
||||
sys.path.insert(0,str(APP))
|
||||
|
||||
import aios_autonomous_control_plane as acp
|
||||
import aios_capability_contract as contract
|
||||
import aios_release_online_agent_worker as worker
|
||||
|
||||
|
||||
def _ready_probes():
|
||||
return {
|
||||
"credential_projection": lambda: True,
|
||||
"trusted_builder": lambda: True,
|
||||
"authenticated_probe": lambda: True,
|
||||
"release_api_health": lambda: True,
|
||||
"worker_liveness": lambda: True,
|
||||
}
|
||||
|
||||
|
||||
def _task():
|
||||
return {
|
||||
"task_id":"r42-autonomous-release-admission-test",
|
||||
"required_capabilities":["release:candidate"],
|
||||
"constraints":{
|
||||
"release_candidate":{
|
||||
"candidate_id":"r42-release-candidate-v2r2c9-test",
|
||||
"source_commit":"1"*40,
|
||||
"version":"V2R2C9",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
}
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
def test_capability_contract_registry_is_closed_and_release_is_versioned():
|
||||
status=contract.validate_capability_contract_registry()
|
||||
assert status["valid"] is True
|
||||
release=contract.capability_contract("release:candidate")
|
||||
assert release is not None
|
||||
assert release.min_version==1
|
||||
assert release.max_version==1
|
||||
assert release.permissions==("prepare_release",)
|
||||
assert "credential_projection" in release.dependencies
|
||||
assert "trusted_builder" in release.dependencies
|
||||
assert "authenticated_probe" in release.health_requirements
|
||||
assert "production_deploy" in release.forbidden
|
||||
assert "runtime_switch" in release.forbidden
|
||||
|
||||
|
||||
def test_unknown_agent_capability_cannot_register_or_match():
|
||||
assert contract.registered_capability("release:future",worker=True) is False
|
||||
assert contract.registered_capability("release:*",worker=True) is False
|
||||
assert contract.has_capabilities(["release:future"],["release:future"]) is False
|
||||
|
||||
|
||||
def test_runtime_admission_requires_all_release_execution_readiness():
|
||||
decision=acp.runtime_admission(
|
||||
worker_capabilities=["release:candidate"],
|
||||
required_capabilities=["release:candidate"],
|
||||
requested_operations=["prepare_release"],
|
||||
probes=_ready_probes(),
|
||||
)
|
||||
assert decision.admitted is True
|
||||
assert decision.state=="EXECUTION_READY"
|
||||
|
||||
|
||||
def test_runtime_admission_blocks_credential_failure_before_execution():
|
||||
probes=_ready_probes()
|
||||
probes["credential_projection"]=lambda:False
|
||||
probes["authenticated_probe"]=lambda:False
|
||||
decision=acp.runtime_admission(
|
||||
worker_capabilities=["release:candidate"],
|
||||
required_capabilities=["release:candidate"],
|
||||
requested_operations=["prepare_release"],
|
||||
probes=probes,
|
||||
)
|
||||
assert decision.admitted is False
|
||||
assert decision.state=="EXECUTION_READY_FAILED"
|
||||
assert decision.failure_class==acp.FailureClass.CREDENTIAL_PROJECTION_ERROR
|
||||
assert decision.repair_level==acp.RepairLevel.AUTO_REPAIR_TASK
|
||||
|
||||
|
||||
def test_forbidden_release_operation_requires_owner_boundary():
|
||||
decision=acp.runtime_admission(
|
||||
worker_capabilities=["release:candidate"],
|
||||
required_capabilities=["release:candidate"],
|
||||
requested_operations=["production_deploy"],
|
||||
probes=_ready_probes(),
|
||||
)
|
||||
assert decision.admitted is False
|
||||
assert decision.failure_class==acp.FailureClass.POLICY_DENIED
|
||||
assert decision.repair_level==acp.RepairLevel.OWNER_APPROVAL
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"error,phase,expected,level,consumes",
|
||||
[
|
||||
("Permission denied: gitops.tokens","release",acp.FailureClass.CREDENTIAL_PROJECTION_ERROR,acp.RepairLevel.AUTO_REPAIR_TASK,False),
|
||||
("HTTP_TRANSPORT_UNAVAILABLE","network",acp.FailureClass.DEPENDENCY_UNAVAILABLE,acp.RepairLevel.AUTO_LOCAL,False),
|
||||
("TASK_CAPABILITY_SCOPE_INVALID","validation",acp.FailureClass.CAPABILITY_CONTRACT_ERROR,acp.RepairLevel.AUTO_REPAIR_TASK,False),
|
||||
("lease lost","heartbeat",acp.FailureClass.LEASE_LOST,acp.RepairLevel.AUTO_LOCAL,False),
|
||||
("business permanent invalid release","release",acp.FailureClass.BUSINESS_PERMANENT,acp.RepairLevel.MANUAL,True),
|
||||
],
|
||||
)
|
||||
def test_failure_classification_drives_repair_policy(error,phase,expected,level,consumes):
|
||||
failure=acp.classify_failure(error,phase=phase)
|
||||
assert failure==expected
|
||||
repair=acp.repair_for_failure(failure)
|
||||
assert repair.level==level
|
||||
assert repair.consumes_business_attempt is consumes
|
||||
|
||||
|
||||
def test_release_worker_does_not_claim_when_admission_is_not_ready(monkeypatch):
|
||||
calls=[]
|
||||
def fake_http(base,method,route,token,body=None,timeout=30):
|
||||
calls.append((base,method,route,body))
|
||||
if route=="/health":
|
||||
return 200,{"status":"ok"}
|
||||
if route=="/createReleaseCandidate" and body=={}:
|
||||
return 500,{"error":"Permission denied: gitops.tokens"}
|
||||
if route.endswith("/claim"):
|
||||
raise AssertionError("claim must not happen before EXECUTION_READY")
|
||||
raise AssertionError(route)
|
||||
|
||||
events=[]
|
||||
monkeypatch.setattr(worker,"http_json",fake_http)
|
||||
monkeypatch.setattr(worker,"audit",lambda event,**fields:events.append((event,fields)))
|
||||
monkeypatch.setattr(worker,"write_liveness",lambda *args,**kwargs:None)
|
||||
|
||||
assert worker.execute_task(
|
||||
_task(),
|
||||
agent_token="a"*32,
|
||||
release_token="r"*32,
|
||||
) is False
|
||||
assert not any(route.endswith("/claim") for _,_,route,_ in calls)
|
||||
blocked=[item for item in events if item[0]=="TASK_ADMISSION_BLOCKED"]
|
||||
assert len(blocked)==1
|
||||
assert blocked[0][1]["admission"]["state"]=="EXECUTION_READY_FAILED"
|
||||
|
||||
|
||||
def test_system_invariant_snapshot_is_degraded_when_dependency_is_missing():
|
||||
probes=_ready_probes()
|
||||
probes["trusted_builder"]=lambda:False
|
||||
snapshot=acp.system_invariant_snapshot(
|
||||
worker_capabilities=["release:candidate"],
|
||||
required_capabilities=["release:candidate"],
|
||||
probes=probes,
|
||||
)
|
||||
assert snapshot["status"]=="DEGRADED"
|
||||
assert snapshot["execution_ready"] is False
|
||||
@@ -0,0 +1,627 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime, timezone
|
||||
import json
|
||||
from pathlib import Path
|
||||
import sqlite3
|
||||
import sys
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
APP=ROOT/"active-runtime/app"
|
||||
sys.path.insert(0,str(APP))
|
||||
|
||||
import aios_durable_owner_authorization as auth
|
||||
|
||||
|
||||
NOW=datetime(2026,8,19,6,0,0,tzinfo=timezone.utc)
|
||||
ACCOUNT="chatgpt-governed-implementer-singlehost"
|
||||
|
||||
|
||||
def write_root_file(path: Path, value: object) -> Path:
|
||||
path.parent.mkdir(parents=True,exist_ok=True)
|
||||
path.write_text(json.dumps(value,sort_keys=True)+"\n",encoding="utf-8")
|
||||
path.chmod(0o600)
|
||||
return path
|
||||
|
||||
|
||||
def policy_value(**overrides):
|
||||
value={
|
||||
"schema":auth.POLICY_SCHEMA,
|
||||
"grant_id":"owner-permanent-preparation-20260819",
|
||||
"owner_identity":"owner-singlehost",
|
||||
"account":ACCOUNT,
|
||||
"capabilities":[
|
||||
"governed-task:project",
|
||||
"release:candidate",
|
||||
],
|
||||
"issued_at":"2026-08-19T05:00:00Z",
|
||||
"not_before":"2026-08-19T05:00:00Z",
|
||||
"persistent_until_revoked":True,
|
||||
"constraints":{
|
||||
"repository":"aios/aios",
|
||||
"allowed_target_prefix":"agent/",
|
||||
"release_version_prefix":"V2R2C",
|
||||
"max_task_risk":"engineering",
|
||||
},
|
||||
}
|
||||
value.update(overrides)
|
||||
return value
|
||||
|
||||
|
||||
def load_policy(tmp_path,monkeypatch,**overrides):
|
||||
path=write_root_file(tmp_path/"policy.json",policy_value(**overrides))
|
||||
revocations=write_root_file(
|
||||
tmp_path/"revocations.json",
|
||||
{"revoked_grant_ids":[]},
|
||||
)
|
||||
monkeypatch.setattr(auth,"validate_root_file",lambda path,**kwargs:path)
|
||||
return auth.DurablePolicy.load(
|
||||
path,
|
||||
revocations,
|
||||
now=NOW,
|
||||
),path,revocations
|
||||
|
||||
|
||||
def governed_request(**overrides):
|
||||
value={
|
||||
"task_id":"r42-owner-auth-projection-test-20260819",
|
||||
"repository":"aios/aios",
|
||||
"base_oid":"8"*40,
|
||||
"expected_branch_oid":"8"*40,
|
||||
"target_branch":"agent/r42-owner-auth-projection-test-20260819",
|
||||
"generation":1,
|
||||
"risk":"engineering",
|
||||
"description":"project existing governed task",
|
||||
"idempotency_key":"binding-owner-auth-projection-v1",
|
||||
"branch_mode":"CREATE_NEW",
|
||||
"priority":100,
|
||||
"max_attempts":3,
|
||||
"require_online_task_projection":True,
|
||||
"online_task":{
|
||||
"goal":"Analyze source read-only.",
|
||||
"instructions":"Use source capture first.",
|
||||
"success_criteria":[
|
||||
{"id":"source","description":"Return source snapshot."}
|
||||
],
|
||||
"constraints":{"repository_mutation":"forbidden"},
|
||||
"required_capabilities":["git.read","git.test"],
|
||||
"priority":100,
|
||||
"max_attempts":3,
|
||||
},
|
||||
"execution":{
|
||||
"action":"source_capture",
|
||||
"adapter":"source_capture",
|
||||
"capability_scope":["git.read","git.test"],
|
||||
"source_signals":["durable authorization projection"],
|
||||
},
|
||||
}
|
||||
value.update(overrides)
|
||||
return value
|
||||
|
||||
|
||||
def release_request(**overrides):
|
||||
value={
|
||||
"candidate_id":"r42-pf1-release-candidate-20260819",
|
||||
"source_commit":"8"*40,
|
||||
"version":"V2R2C2",
|
||||
}
|
||||
value.update(overrides)
|
||||
return value
|
||||
|
||||
|
||||
class FakeUpstream:
|
||||
def __init__(self):
|
||||
self.project_calls=[]
|
||||
|
||||
def project_governed_task(self,body):
|
||||
self.project_calls.append(dict(body))
|
||||
return 200,{
|
||||
"task_id":body["task_id"],
|
||||
"job_id":"job-"+"a"*32,
|
||||
"state":"BOUND",
|
||||
}
|
||||
|
||||
|
||||
def broker(tmp_path,monkeypatch):
|
||||
policy,path,revocations=load_policy(tmp_path,monkeypatch)
|
||||
store=auth.ReceiptStore(tmp_path/"receipts.sqlite3")
|
||||
upstream=FakeUpstream()
|
||||
instance=auth.Broker(
|
||||
policy_path=path,
|
||||
revocation_path=revocations,
|
||||
store=store,
|
||||
upstream=upstream,
|
||||
expected_account=ACCOUNT,
|
||||
)
|
||||
monkeypatch.setattr(
|
||||
auth.DurablePolicy,
|
||||
"load",
|
||||
classmethod(lambda cls,*args,**kwargs:policy),
|
||||
)
|
||||
return instance,upstream,store,policy
|
||||
|
||||
|
||||
def test_allowed_persistent_capabilities_are_only_preparation_scopes():
|
||||
assert auth.ALLOWED_PERSISTENT_CAPABILITIES=={
|
||||
"governed-task:project",
|
||||
"release:candidate",
|
||||
}
|
||||
assert auth.ALLOWED_PERSISTENT_CAPABILITIES.isdisjoint(
|
||||
auth.FORBIDDEN_PERSISTENT_CAPABILITIES
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"capability",
|
||||
sorted(auth.FORBIDDEN_PERSISTENT_CAPABILITIES),
|
||||
)
|
||||
def test_policy_rejects_high_risk_persistent_capability(
|
||||
tmp_path,monkeypatch,capability
|
||||
):
|
||||
bad=policy_value(
|
||||
capabilities=["governed-task:project",capability]
|
||||
)
|
||||
path=write_root_file(tmp_path/"policy.json",bad)
|
||||
revocations=write_root_file(
|
||||
tmp_path/"revocations.json",
|
||||
{"revoked_grant_ids":[]},
|
||||
)
|
||||
monkeypatch.setattr(auth,"validate_root_file",lambda path,**kwargs:path)
|
||||
with pytest.raises(
|
||||
auth.AuthorizationError,
|
||||
match="POLICY_CAPABILITY_NOT_PERSISTABLE",
|
||||
):
|
||||
auth.DurablePolicy.load(path,revocations,now=NOW)
|
||||
|
||||
|
||||
def test_policy_is_persistent_until_revoked(tmp_path,monkeypatch):
|
||||
policy,_,_=load_policy(tmp_path,monkeypatch)
|
||||
assert policy.persistent_until_revoked is True
|
||||
assert policy.account==ACCOUNT
|
||||
assert policy.capabilities=={
|
||||
"governed-task:project",
|
||||
"release:candidate",
|
||||
}
|
||||
|
||||
|
||||
def test_revocation_fails_closed(tmp_path,monkeypatch):
|
||||
path=write_root_file(tmp_path/"policy.json",policy_value())
|
||||
revocations=write_root_file(
|
||||
tmp_path/"revocations.json",
|
||||
{"revoked_grant_ids":["owner-permanent-preparation-20260819"]},
|
||||
)
|
||||
monkeypatch.setattr(auth,"validate_root_file",lambda path,**kwargs:path)
|
||||
with pytest.raises(auth.AuthorizationError,match="OWNER_GRANT_REVOKED"):
|
||||
auth.DurablePolicy.load(path,revocations,now=NOW)
|
||||
|
||||
|
||||
def test_future_policy_is_not_active(tmp_path,monkeypatch):
|
||||
with pytest.raises(auth.AuthorizationError,match="POLICY_NOT_ACTIVE"):
|
||||
load_policy(
|
||||
tmp_path,
|
||||
monkeypatch,
|
||||
issued_at="2026-08-20T05:00:00Z",
|
||||
not_before="2026-08-20T05:00:00Z",
|
||||
)
|
||||
|
||||
|
||||
def test_governed_projection_request_is_tightly_constrained(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
policy,_,_=load_policy(tmp_path,monkeypatch)
|
||||
decision=policy.authorize(
|
||||
"governed-task:project",
|
||||
governed_request(),
|
||||
)
|
||||
assert decision["decision"]=="ALLOW"
|
||||
assert decision["persistent_until_revoked"] is True
|
||||
|
||||
bad_caps=governed_request()
|
||||
bad_caps["online_task"]={
|
||||
**bad_caps["online_task"],
|
||||
"required_capabilities":["release:candidate"],
|
||||
}
|
||||
bad_action=governed_request()
|
||||
bad_action["execution"]={
|
||||
**bad_action["execution"],
|
||||
"action":"deterministic",
|
||||
"adapter":"deterministic",
|
||||
}
|
||||
bad_scope=governed_request()
|
||||
bad_scope["execution"]={
|
||||
**bad_scope["execution"],
|
||||
"capability_scope":["git.read"],
|
||||
}
|
||||
for bad in (
|
||||
governed_request(repository="other/repo"),
|
||||
governed_request(target_branch="main"),
|
||||
governed_request(generation=2),
|
||||
governed_request(risk="high"),
|
||||
governed_request(expected_branch_oid="9"*40),
|
||||
governed_request(branch_mode="CONTINUE_EXISTING"),
|
||||
governed_request(priority=101),
|
||||
governed_request(max_attempts=4),
|
||||
governed_request(require_online_task_projection=False),
|
||||
bad_caps,
|
||||
bad_action,
|
||||
bad_scope,
|
||||
):
|
||||
with pytest.raises(auth.AuthorizationError):
|
||||
policy.authorize("governed-task:project",bad)
|
||||
|
||||
|
||||
|
||||
def test_write_capable_projection_is_allowed_only_as_source_capture_placeholder(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
policy,_,_=load_policy(tmp_path,monkeypatch)
|
||||
request=governed_request()
|
||||
request["online_task"]={
|
||||
**request["online_task"],
|
||||
"required_capabilities":[
|
||||
"git.read",
|
||||
"git.test",
|
||||
"governed.write.branch",
|
||||
"governed.write.file",
|
||||
],
|
||||
}
|
||||
request["execution"]={
|
||||
**request["execution"],
|
||||
"capability_scope":[
|
||||
"git.read",
|
||||
"git.test",
|
||||
"governed.write.branch",
|
||||
"governed.write.file",
|
||||
],
|
||||
}
|
||||
decision=policy.authorize("governed-task:project",request)
|
||||
assert decision["decision"]=="ALLOW"
|
||||
assert request["execution"]["action"]=="source_capture"
|
||||
|
||||
def test_release_candidate_request_is_tightly_constrained(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
policy,_,_=load_policy(tmp_path,monkeypatch)
|
||||
decision=policy.authorize("release:candidate",release_request())
|
||||
assert decision["decision"]=="ALLOW"
|
||||
|
||||
with pytest.raises(
|
||||
auth.AuthorizationError,
|
||||
match="RELEASE_VERSION_NOT_AUTHORIZED",
|
||||
):
|
||||
policy.authorize(
|
||||
"release:candidate",
|
||||
release_request(version="V3R1C1"),
|
||||
)
|
||||
with pytest.raises(
|
||||
auth.AuthorizationError,
|
||||
match="RELEASE_CANDIDATE_FIELDS_INVALID",
|
||||
):
|
||||
policy.authorize(
|
||||
"release:candidate",
|
||||
{**release_request(),"execute":True},
|
||||
)
|
||||
|
||||
|
||||
def test_governed_task_projection_executes_canonical_upstream(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,upstream,store,_=broker(tmp_path,monkeypatch)
|
||||
request=governed_request()
|
||||
response=instance.execute(
|
||||
"governed-task:project",
|
||||
request,
|
||||
idempotency_key="owner-auth-projection-operation-0001",
|
||||
)
|
||||
assert len(upstream.project_calls)==1
|
||||
assert upstream.project_calls[0]==request
|
||||
assert response["operation_state"]=="EXECUTED"
|
||||
assert response["external_capability_projection_required"] is False
|
||||
assert response["separate_owner_release_approval_still_required"] is True
|
||||
|
||||
|
||||
def test_release_candidate_authorization_does_not_invent_builder_upstream(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,upstream,store,_=broker(tmp_path,monkeypatch)
|
||||
response=instance.execute(
|
||||
"release:candidate",
|
||||
release_request(),
|
||||
idempotency_key="owner-auth-release-candidate-operation-0001",
|
||||
)
|
||||
assert upstream.project_calls==[]
|
||||
assert response["operation_state"]==(
|
||||
"AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION"
|
||||
)
|
||||
assert response["upstream_status"] is None
|
||||
assert response["upstream_response"] is None
|
||||
assert response["external_capability_projection_required"] is True
|
||||
assert response["projection_id"].startswith("projection-")
|
||||
assert response["projection_state"]=="PENDING_EXTERNAL_CONNECTOR"
|
||||
assert response["separate_owner_release_approval_still_required"] is True
|
||||
|
||||
|
||||
|
||||
def test_release_authorization_creates_durable_projection_outbox(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,policy=broker(tmp_path,monkeypatch)
|
||||
request=release_request()
|
||||
response=instance.execute(
|
||||
"release:candidate",
|
||||
request,
|
||||
idempotency_key="owner-auth-release-outbox-0001",
|
||||
)
|
||||
projection=store.get_projection(response["projection_id"])
|
||||
assert projection is not None
|
||||
assert projection["schema"]==auth.PROJECTION_SCHEMA
|
||||
assert projection["state"]=="PENDING_EXTERNAL_CONNECTOR"
|
||||
assert projection["grant_id"]==policy.grant_id
|
||||
assert projection["account"]==ACCOUNT
|
||||
assert projection["capability"]=="release:candidate"
|
||||
assert projection["request"]==request
|
||||
assert projection["separate_owner_release_approval_still_required"] is True
|
||||
|
||||
listed=instance.projections()
|
||||
assert listed["schema"]==auth.PROJECTION_LIST_SCHEMA
|
||||
assert listed["count"]==1
|
||||
assert listed["items"][0]["projection_id"]==response["projection_id"]
|
||||
assert listed["external_connector_ack_supported"] is False
|
||||
|
||||
|
||||
|
||||
def test_existing_pre_outbox_release_receipt_is_backfilled_on_exact_retry(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,policy=broker(tmp_path,monkeypatch)
|
||||
request=release_request()
|
||||
decision=policy.authorize("release:candidate",request)
|
||||
key="owner-auth-release-pre-outbox-0001"
|
||||
legacy_response={
|
||||
"schema":auth.RECEIPT_SCHEMA,
|
||||
"grant_id":policy.grant_id,
|
||||
"owner_identity":policy.owner_identity,
|
||||
"account":policy.account,
|
||||
"capability":"release:candidate",
|
||||
"policy_sha256":policy.policy_sha256,
|
||||
"request_sha256":decision["request_sha256"],
|
||||
"idempotency_key":key,
|
||||
"operation_state":"AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION",
|
||||
"upstream_status":None,
|
||||
"upstream_response":None,
|
||||
"external_capability_projection_required":True,
|
||||
"separate_owner_release_approval_still_required":True,
|
||||
}
|
||||
store.put(
|
||||
key,
|
||||
"release:candidate",
|
||||
decision["request_sha256"],
|
||||
legacy_response,
|
||||
)
|
||||
response=instance.execute(
|
||||
"release:candidate",
|
||||
request,
|
||||
idempotency_key=key,
|
||||
)
|
||||
assert response["projection_id"].startswith("projection-")
|
||||
assert response["projection_state"]=="PENDING_EXTERNAL_CONNECTOR"
|
||||
assert store.get_projection(response["projection_id"]) is not None
|
||||
persisted=store.get(key)["response"]
|
||||
assert persisted["projection_id"]==response["projection_id"]
|
||||
|
||||
def test_projection_id_is_deterministic_and_exact_retry_does_not_duplicate(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,_=broker(tmp_path,monkeypatch)
|
||||
key="owner-auth-release-outbox-0002"
|
||||
first=instance.execute("release:candidate",release_request(),idempotency_key=key)
|
||||
second=instance.execute("release:candidate",release_request(),idempotency_key=key)
|
||||
assert second==first
|
||||
con=sqlite3.connect(store.path)
|
||||
rows=con.execute("select projection_id from projection_outbox").fetchall()
|
||||
assert rows==[(first["projection_id"],)]
|
||||
|
||||
|
||||
def test_projection_enqueue_recovers_crash_between_outbox_and_receipt(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,policy=broker(tmp_path,monkeypatch)
|
||||
request=release_request()
|
||||
decision=policy.authorize("release:candidate",request)
|
||||
key="owner-auth-release-outbox-crash-0001"
|
||||
first=store.enqueue_projection(
|
||||
idempotency_key=key,
|
||||
grant_id=policy.grant_id,
|
||||
account=policy.account,
|
||||
capability="release:candidate",
|
||||
policy_sha256=policy.policy_sha256,
|
||||
request_sha256=decision["request_sha256"],
|
||||
request=request,
|
||||
)
|
||||
# Simulate a later retry after the process crashed before receipt creation.
|
||||
second=store.enqueue_projection(
|
||||
idempotency_key=key,
|
||||
grant_id=policy.grant_id,
|
||||
account=policy.account,
|
||||
capability="release:candidate",
|
||||
policy_sha256=policy.policy_sha256,
|
||||
request_sha256=decision["request_sha256"],
|
||||
request=request,
|
||||
)
|
||||
assert second==first
|
||||
|
||||
|
||||
def test_projection_idempotency_conflict_is_fail_closed(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,policy=broker(tmp_path,monkeypatch)
|
||||
first=release_request()
|
||||
first_decision=policy.authorize("release:candidate",first)
|
||||
key="owner-auth-release-outbox-conflict-0001"
|
||||
store.enqueue_projection(
|
||||
idempotency_key=key,
|
||||
grant_id=policy.grant_id,
|
||||
account=policy.account,
|
||||
capability="release:candidate",
|
||||
policy_sha256=policy.policy_sha256,
|
||||
request_sha256=first_decision["request_sha256"],
|
||||
request=first,
|
||||
)
|
||||
second=release_request(version="V2R2C3")
|
||||
second_decision=policy.authorize("release:candidate",second)
|
||||
with pytest.raises(auth.AuthorizationError,match="PROJECTION_IDEMPOTENCY_CONFLICT"):
|
||||
store.enqueue_projection(
|
||||
idempotency_key=key,
|
||||
grant_id=policy.grant_id,
|
||||
account=policy.account,
|
||||
capability="release:candidate",
|
||||
policy_sha256=policy.policy_sha256,
|
||||
request_sha256=second_decision["request_sha256"],
|
||||
request=second,
|
||||
)
|
||||
|
||||
|
||||
def test_revocation_blocks_projection_enumeration(tmp_path,monkeypatch):
|
||||
instance,_,store,policy=broker(tmp_path,monkeypatch)
|
||||
response=instance.execute(
|
||||
"release:candidate",
|
||||
release_request(),
|
||||
idempotency_key="owner-auth-release-outbox-revoke-0001",
|
||||
)
|
||||
# Restore real policy loading and write a revocation for this grant.
|
||||
monkeypatch.undo()
|
||||
revocations=instance.revocation_path
|
||||
revocations.write_text(
|
||||
json.dumps({"revoked_grant_ids":[policy.grant_id]})+"\n",
|
||||
encoding="utf-8",
|
||||
)
|
||||
revocations.chmod(0o600)
|
||||
monkeypatch.setattr(auth,"validate_root_file",lambda path,**kwargs:path)
|
||||
with pytest.raises(auth.AuthorizationError,match="OWNER_GRANT_REVOKED"):
|
||||
instance.projections()
|
||||
# Evidence is retained; revocation does not erase history.
|
||||
assert store.get_projection(response["projection_id"]) is not None
|
||||
|
||||
|
||||
def test_projection_outbox_contains_no_credentials(tmp_path,monkeypatch):
|
||||
instance,_,store,_=broker(tmp_path,monkeypatch)
|
||||
instance.execute(
|
||||
"release:candidate",
|
||||
release_request(),
|
||||
idempotency_key="owner-auth-release-outbox-schema-0001",
|
||||
)
|
||||
con=sqlite3.connect(store.path)
|
||||
columns=[row[1] for row in con.execute("pragma table_info(projection_outbox)")]
|
||||
joined=" ".join(columns).lower()
|
||||
for forbidden in ("token","credential","private_key","secret"):
|
||||
assert forbidden not in joined
|
||||
payload=con.execute("select payload_json from projection_outbox").fetchone()[0]
|
||||
assert "Bearer " not in payload
|
||||
assert "submitOwnerApproval" not in payload
|
||||
assert "startDeployment" not in payload
|
||||
|
||||
def test_exact_retry_returns_same_receipt_without_second_projection(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,upstream,_,_=broker(tmp_path,monkeypatch)
|
||||
request=governed_request()
|
||||
first=instance.execute(
|
||||
"governed-task:project",
|
||||
request,
|
||||
idempotency_key="owner-auth-projection-operation-0002",
|
||||
)
|
||||
second=instance.execute(
|
||||
"governed-task:project",
|
||||
request,
|
||||
idempotency_key="owner-auth-projection-operation-0002",
|
||||
)
|
||||
assert second==first
|
||||
assert len(upstream.project_calls)==1
|
||||
|
||||
|
||||
def test_idempotency_conflict_fails_closed(tmp_path,monkeypatch):
|
||||
instance,_,_,_=broker(tmp_path,monkeypatch)
|
||||
instance.execute(
|
||||
"release:candidate",
|
||||
release_request(),
|
||||
idempotency_key="owner-auth-release-operation-0002",
|
||||
)
|
||||
with pytest.raises(auth.AuthorizationError,match="IDEMPOTENCY_CONFLICT"):
|
||||
instance.execute(
|
||||
"release:candidate",
|
||||
release_request(version="V2R2C3"),
|
||||
idempotency_key="owner-auth-release-operation-0002",
|
||||
)
|
||||
|
||||
|
||||
def test_receipt_store_contains_no_owner_policy_or_credentials(
|
||||
tmp_path,monkeypatch
|
||||
):
|
||||
instance,_,store,_=broker(tmp_path,monkeypatch)
|
||||
instance.execute(
|
||||
"release:candidate",
|
||||
release_request(),
|
||||
idempotency_key="owner-auth-release-operation-0003",
|
||||
)
|
||||
con=sqlite3.connect(store.path)
|
||||
columns=[row[1] for row in con.execute("pragma table_info(receipts)")]
|
||||
joined=" ".join(columns).lower()
|
||||
assert "token" not in joined
|
||||
assert "credential" not in joined
|
||||
assert "policy_json" not in columns
|
||||
|
||||
|
||||
def test_systemd_broker_uses_unix_socket_and_only_host_control_credential():
|
||||
unit=(
|
||||
ROOT/"deploy/systemd/aios-owner-durable-authorization-broker.service"
|
||||
).read_text()
|
||||
assert "LoadCredential=agent_task_control_token:" in unit
|
||||
assert "release_candidate_token" not in unit
|
||||
assert "18947" not in unit
|
||||
assert "AIOS_RELEASE_CANDIDATE_API_URL" not in unit
|
||||
assert "AIOS_OWNER_AUTH_BROKER_SOCKET=%t/aios-owner-authorization-broker/broker.sock" in unit
|
||||
assert "Group=aios-executor-exchange" in unit
|
||||
assert "RestrictAddressFamilies=AF_UNIX AF_INET" in unit
|
||||
assert "IPAddressDeny=any" in unit
|
||||
assert "IPAddressAllow=127.0.0.0/8" in unit
|
||||
assert "InaccessiblePaths=/data/aios-gitops /home" in unit
|
||||
|
||||
|
||||
def test_source_contains_no_release_builder_endpoint_or_high_risk_execution():
|
||||
source=(
|
||||
ROOT/"active-runtime/app/aios_durable_owner_authorization.py"
|
||||
).read_text()
|
||||
assert "/createReleaseCandidate" not in source
|
||||
assert "18947" not in source
|
||||
assert "ThreadingHTTPServer" not in source
|
||||
assert "ThreadingUnixStreamServer" in source
|
||||
assert "startDeployment" in source
|
||||
assert "tag_publish" in source
|
||||
assert "baseline_registry_finalize" in source
|
||||
assert "runtime_switch" in source
|
||||
assert "AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION" in source
|
||||
assert "PENDING_EXTERNAL_CONNECTOR" in source
|
||||
assert "projection_outbox" in source
|
||||
assert "external_connector_ack_supported" in source
|
||||
assert "SAFE_PROJECTED_TASK_CAPABILITIES" in source
|
||||
assert "INITIAL_EXECUTION_ACTION_NOT_AUTHORIZED" in source
|
||||
|
||||
|
||||
|
||||
def test_docs_avoid_new_broker_bearer_secret():
|
||||
doc=(
|
||||
ROOT/"docs/superpowers/specs/2026-08-19-durable-owner-authorization.md"
|
||||
).read_text()
|
||||
assert "AF_UNIX" in doc
|
||||
assert "avoids introducing another long-lived bearer secret" in doc
|
||||
|
||||
def test_docs_preserve_separate_owner_release_approval():
|
||||
doc=(
|
||||
ROOT/"docs/superpowers/specs/2026-08-19-durable-owner-authorization.md"
|
||||
).read_text()
|
||||
assert "persistent until explicitly revoked" in doc.lower()
|
||||
assert "Owner signature" in doc
|
||||
assert "AUTHORIZED_PENDING_EXTERNAL_CAPABILITY_PROJECTION" in doc
|
||||
assert "PENDING_EXTERNAL_CONNECTOR" in doc
|
||||
@@ -0,0 +1,112 @@
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
import yaml
|
||||
|
||||
from aios_chatgpt_bridge_v2 import OPERATION_REQUIRED_FIELDS
|
||||
from aios_chatgpt_bridge_v2_server import BridgeHandler, _HTTPError, ROUTES
|
||||
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
OPENAPI = ROOT / "docs" / "gitops" / "openapi-chatgpt-actions.yaml"
|
||||
REQUIRED = {
|
||||
"task_id",
|
||||
"repository",
|
||||
"base_oid",
|
||||
"target_branch",
|
||||
"risk",
|
||||
"description",
|
||||
"idempotency_key",
|
||||
}
|
||||
OPTIONAL_CONTINUATION = {
|
||||
"expected_branch_oid",
|
||||
"generation",
|
||||
"parent_task_id",
|
||||
"parent_job_id",
|
||||
"priority",
|
||||
}
|
||||
|
||||
|
||||
def _document():
|
||||
return yaml.safe_load(OPENAPI.read_text(encoding="utf-8"))
|
||||
|
||||
|
||||
def _create_request():
|
||||
return {
|
||||
"task_id": "continuation-contract-test",
|
||||
"repository": "aios/aios",
|
||||
"base_oid": "a" * 40,
|
||||
"target_branch": "agent/continuation-contract-test",
|
||||
"risk": "engineering",
|
||||
"description": "contract test",
|
||||
"idempotency_key": "continuation-contract-test-v1",
|
||||
}
|
||||
|
||||
|
||||
def test_create_governed_task_openapi_contract_exposes_continuation_fields():
|
||||
document = _document()
|
||||
operation = document["paths"]["/createGovernedTask"]["post"]
|
||||
schema_ref = operation["requestBody"]["content"]["application/json"]["schema"]
|
||||
schema = document["components"]["schemas"][schema_ref["$ref"].rsplit("/", 1)[-1]]
|
||||
|
||||
assert operation["operationId"] == "createGovernedTask"
|
||||
assert operation["security"] == [{"bearerAuth": []}]
|
||||
assert operation["x-openai-isConsequential"] is True
|
||||
assert set(schema["required"]) == REQUIRED
|
||||
assert schema["additionalProperties"] is False
|
||||
assert OPTIONAL_CONTINUATION <= set(schema["properties"])
|
||||
assert not OPTIONAL_CONTINUATION.intersection(schema["required"])
|
||||
assert schema["properties"]["repository"]["const"] == "aios/aios"
|
||||
assert schema["properties"]["generation"]["minimum"] == 1
|
||||
assert schema["properties"]["priority"] == {
|
||||
"type": "integer",
|
||||
"minimum": -100,
|
||||
"maximum": 100,
|
||||
}
|
||||
|
||||
|
||||
def test_create_governed_task_route_and_required_fields_match_bridge():
|
||||
assert ROUTES["/v2/governed-task"] == "createGovernedTask"
|
||||
assert tuple(OPERATION_REQUIRED_FIELDS["createGovernedTask"]) == (
|
||||
"task_id",
|
||||
"repository",
|
||||
"base_oid",
|
||||
"target_branch",
|
||||
"risk",
|
||||
"description",
|
||||
"idempotency_key",
|
||||
)
|
||||
|
||||
|
||||
def test_bridge_execution_plan_and_envelope_routes_remain_exposed():
|
||||
assert ROUTES["/v2/execution-envelope"] == "projectOnlineAgentExecution"
|
||||
assert ROUTES["/v2/execution-plan"] == "registerOnlineAgentExecutionPlan"
|
||||
|
||||
|
||||
def test_bridge_allowlist_accepts_parent_pair_and_priority_but_rejects_unknown():
|
||||
handler = BridgeHandler.__new__(BridgeHandler)
|
||||
body = _create_request()
|
||||
body.update(
|
||||
{
|
||||
"expected_branch_oid": "a" * 40,
|
||||
"generation": 2,
|
||||
"parent_task_id": "continuation-parent-test",
|
||||
"parent_job_id": "job-continuation-parent-test",
|
||||
"priority": 100,
|
||||
}
|
||||
)
|
||||
handler._allowed("createGovernedTask", body)
|
||||
|
||||
invalid = dict(body)
|
||||
invalid["unexpected_field"] = True
|
||||
with pytest.raises(_HTTPError, match="UNKNOWN_FIELDS"):
|
||||
handler._allowed("createGovernedTask", invalid)
|
||||
|
||||
|
||||
def test_openapi_does_not_add_credential_fields_to_create_contract():
|
||||
document = _document()
|
||||
properties = set(
|
||||
document["components"]["schemas"]["CreateGovernedTaskRequest"]["properties"]
|
||||
)
|
||||
forbidden = {"authorization", "bearer", "token", "credential", "secret", "password", "api_key"}
|
||||
assert properties.isdisjoint(forbidden)
|
||||
@@ -0,0 +1,83 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import sys
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0,str(ROOT/"active-runtime/app"))
|
||||
|
||||
from aios_capability_contract import ( # noqa: E402
|
||||
has_capabilities,
|
||||
normalize_required_capabilities,
|
||||
normalize_worker_capabilities,
|
||||
normalized_capability_matrix,
|
||||
required_capabilities_subset,
|
||||
)
|
||||
|
||||
|
||||
def test_release_candidate_is_known_to_task_and_worker_vocabularies():
|
||||
capability="release:candidate"
|
||||
assert normalize_required_capabilities([capability])==frozenset({capability})
|
||||
assert normalize_worker_capabilities([capability])==frozenset({capability})
|
||||
assert has_capabilities([capability],[capability]) is True
|
||||
assert normalized_capability_matrix([capability],[capability])=={
|
||||
"worker_effective":["release:candidate"],
|
||||
"task_required":["release:candidate"],
|
||||
"eligible":True,
|
||||
}
|
||||
|
||||
|
||||
def test_release_candidate_capability_does_not_imply_high_risk_release_operations():
|
||||
worker=["release:candidate"]
|
||||
for forbidden in (
|
||||
"submitOwnerApproval",
|
||||
"startDeployment",
|
||||
"tag_publish",
|
||||
"baseline_registry_finalize",
|
||||
"runtime_switch",
|
||||
"deploy.production",
|
||||
):
|
||||
assert has_capabilities(worker,[forbidden]) is False
|
||||
|
||||
|
||||
def test_release_wildcards_and_nearby_unknown_names_remain_fail_closed():
|
||||
assert has_capabilities(["release:candidate"],["release:*"]) is False
|
||||
assert has_capabilities(["release:candidate"],["release:candidate.extra"]) is False
|
||||
assert has_capabilities(["release:*"],["release:candidate"]) is False
|
||||
|
||||
|
||||
def test_release_candidate_delegation_subset_is_exact():
|
||||
assert required_capabilities_subset(
|
||||
["release:candidate"],
|
||||
["release:candidate"],
|
||||
) is True
|
||||
assert required_capabilities_subset(
|
||||
["release:candidate"],
|
||||
["startDeployment"],
|
||||
) is False
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"worker,required",
|
||||
[
|
||||
(["agent.runtime.hermes"],["agent.runtime.hermes"]),
|
||||
(["agent.runtime.openclaw"],["agent.runtime.openclaw"]),
|
||||
(["git.read"],["repository_read"]),
|
||||
(["git.test"],["test_execution"]),
|
||||
(
|
||||
["governed.write.branch","governed.write.file"],
|
||||
["governed.git.write"],
|
||||
),
|
||||
],
|
||||
)
|
||||
def test_existing_known_capability_semantics_are_non_regressed(worker,required):
|
||||
assert has_capabilities(worker,required) is True
|
||||
|
||||
|
||||
def test_release_candidate_worker_does_not_match_empty_or_unrelated_task_by_escalation():
|
||||
assert has_capabilities(["release:candidate"],[]) is True
|
||||
assert has_capabilities(["release:candidate"],["git.read"]) is False
|
||||
assert has_capabilities(["release:candidate"],["agent.runtime.hermes"]) is False
|
||||
@@ -0,0 +1,95 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import sys
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0,str(ROOT/"active-runtime/app"))
|
||||
|
||||
from aios_release_lineage_contract import ( # noqa: E402
|
||||
FIXED_V2R2C2,
|
||||
ReleaseLineageError,
|
||||
validate_fixed_v2r2c2_task,
|
||||
)
|
||||
|
||||
|
||||
def fixed_task(source=FIXED_V2R2C2.source_commit, *, status="FAILED", attempt=1, max_attempts=1, revision=4):
|
||||
return {
|
||||
"task_id":FIXED_V2R2C2.task_id,
|
||||
"status":status,
|
||||
"attempt":attempt,
|
||||
"max_attempts":max_attempts,
|
||||
"revision":revision,
|
||||
"required_capabilities":["release:candidate"],
|
||||
"constraints":{
|
||||
"release_candidate":{
|
||||
"candidate_id":FIXED_V2R2C2.candidate_id,
|
||||
"source_commit":source,
|
||||
"version":FIXED_V2R2C2.version,
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
}
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
def test_fixed_v2r2c2_lineage_is_the_original_release_worker_candidate():
|
||||
assert FIXED_V2R2C2.source_commit=="71e736b05fab13efc06af7a4fe2dc022c1cf9a85"
|
||||
result=validate_fixed_v2r2c2_task(fixed_task())
|
||||
assert result["status"]=="VERIFIED"
|
||||
assert result["lineage"]["version"]=="V2R2C2"
|
||||
|
||||
|
||||
def test_later_capability_remediation_is_not_retroactively_v2r2c2_source():
|
||||
with pytest.raises(ReleaseLineageError,match="FIXED_RELEASE_LINEAGE_DRIFT"):
|
||||
validate_fixed_v2r2c2_task(
|
||||
fixed_task("17deb91ca7a187dc81766945b1174e854f86bd25")
|
||||
)
|
||||
|
||||
|
||||
def test_acp_engineering_candidate_is_not_retroactively_v2r2c2_source():
|
||||
with pytest.raises(ReleaseLineageError,match="FIXED_RELEASE_LINEAGE_DRIFT"):
|
||||
validate_fixed_v2r2c2_task(
|
||||
fixed_task("b4b4416c93fedbe451d655d49e28f915b8fd38e6")
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"status,attempt,max_attempts,revision",
|
||||
[
|
||||
("FAILED",1,1,4),
|
||||
("QUEUED",1,2,5),
|
||||
("CLAIMED",2,2,6),
|
||||
("VERIFYING",2,2,9),
|
||||
("SUCCEEDED",2,2,10),
|
||||
],
|
||||
)
|
||||
def test_lineage_contract_does_not_confuse_lifecycle_state_with_release_identity(
|
||||
status,attempt,max_attempts,revision
|
||||
):
|
||||
result=validate_fixed_v2r2c2_task(
|
||||
fixed_task(
|
||||
status=status,
|
||||
attempt=attempt,
|
||||
max_attempts=max_attempts,
|
||||
revision=revision,
|
||||
)
|
||||
)
|
||||
assert result["status"]=="VERIFIED"
|
||||
assert result["task_status"]==status
|
||||
assert result["attempt"]==attempt
|
||||
assert result["max_attempts"]==max_attempts
|
||||
assert result["revision"]==revision
|
||||
|
||||
|
||||
def test_capability_and_task_identity_are_bound():
|
||||
value=fixed_task()
|
||||
value["required_capabilities"]=[]
|
||||
with pytest.raises(ReleaseLineageError,match="FIXED_TASK_CAPABILITY_DRIFT"):
|
||||
validate_fixed_v2r2c2_task(value)
|
||||
|
||||
value=fixed_task()
|
||||
value["task_id"]="replacement-v2r2c2"
|
||||
with pytest.raises(ReleaseLineageError,match="FIXED_TASK_ID_DRIFT"):
|
||||
validate_fixed_v2r2c2_task(value)
|
||||
@@ -0,0 +1,226 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
from pathlib import Path
|
||||
import sys
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
APP=ROOT/"active-runtime/app"
|
||||
sys.path.insert(0,str(APP))
|
||||
|
||||
import aios_release_online_agent_worker as worker
|
||||
|
||||
|
||||
def task(**overrides):
|
||||
value={
|
||||
"task_id":"r42-release-candidate-live-acceptance-20260819",
|
||||
"required_capabilities":["release:candidate"],
|
||||
"constraints":{
|
||||
"release_candidate":{
|
||||
"candidate_id":"r42-release-candidate-v2r2c2-20260819",
|
||||
"source_commit":"5"*40,
|
||||
"version":"V2R2C2",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
}
|
||||
}
|
||||
}
|
||||
value.update(overrides)
|
||||
return value
|
||||
|
||||
|
||||
def test_release_request_accepts_only_exact_release_capability():
|
||||
req=worker.ReleaseRequest.from_task(task())
|
||||
assert req.candidate_id=="r42-release-candidate-v2r2c2-20260819"
|
||||
assert req.source_commit=="5"*40
|
||||
assert req.version=="V2R2C2"
|
||||
assert req.projection_id=="projection-"+"a"*32
|
||||
|
||||
with pytest.raises(worker.ReleaseWorkerError,match="TASK_CAPABILITY_SCOPE_INVALID"):
|
||||
worker.ReleaseRequest.from_task(task(required_capabilities=["release:candidate","git.read"]))
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"payload",
|
||||
[
|
||||
{
|
||||
"candidate_id":"r42-release-candidate-v2r2c2-20260819",
|
||||
"source_commit":"short",
|
||||
"version":"V2R2C2",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
},
|
||||
{
|
||||
"candidate_id":"bad",
|
||||
"source_commit":"5"*40,
|
||||
"version":"V2R2C2",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
},
|
||||
{
|
||||
"candidate_id":"r42-release-candidate-v2r2c2-20260819",
|
||||
"source_commit":"5"*40,
|
||||
"version":"bad",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
},
|
||||
{
|
||||
"candidate_id":"r42-release-candidate-v2r2c2-20260819",
|
||||
"source_commit":"5"*40,
|
||||
"version":"V2R2C2",
|
||||
"projection_id":"bad",
|
||||
},
|
||||
{
|
||||
"candidate_id":"r42-release-candidate-v2r2c2-20260819",
|
||||
"source_commit":"5"*40,
|
||||
"version":"V2R2C2",
|
||||
"projection_id":"projection-"+"a"*32,
|
||||
"execute":True,
|
||||
},
|
||||
],
|
||||
)
|
||||
def test_release_request_rejects_invalid_or_extra_fields(payload):
|
||||
value=task()
|
||||
value["constraints"]={"release_candidate":payload}
|
||||
with pytest.raises(worker.ReleaseWorkerError):
|
||||
worker.ReleaseRequest.from_task(value)
|
||||
|
||||
|
||||
def test_result_payload_preserves_high_risk_denials():
|
||||
release=worker.ReleaseRequest.from_task(task())
|
||||
result=worker.result_payload(
|
||||
"r42-release-candidate-live-acceptance-20260819",
|
||||
release,
|
||||
{"candidate_id":release.candidate_id,"status":"BUILT"},
|
||||
)
|
||||
assert result["status"]=="SUCCEEDED"
|
||||
assert result["classification"]=="RELEASE_CANDIDATE_CREATED"
|
||||
assert result["writer_lease"]=="NONE"
|
||||
assert result["changed_files"]==[]
|
||||
assert result["direct_main_write"] is False
|
||||
assert result["production_write"] is False
|
||||
assert result["tag_write"] is False
|
||||
assert result["baseline_write"] is False
|
||||
assert result["runtime_switch"] is False
|
||||
assert result["owner_approval_submitted"] is False
|
||||
assert result["credential_material_exposed"] is False
|
||||
|
||||
|
||||
def test_result_payload_rejects_builder_candidate_mismatch():
|
||||
release=worker.ReleaseRequest.from_task(task())
|
||||
with pytest.raises(worker.ReleaseWorkerError,match="BUILDER_CANDIDATE_BINDING_INVALID"):
|
||||
worker.result_payload(
|
||||
"r42-release-candidate-live-acceptance-20260819",
|
||||
release,
|
||||
{"candidate_id":"other-release-candidate"},
|
||||
)
|
||||
|
||||
|
||||
def test_worker_calls_only_fixed_release_candidate_route(monkeypatch):
|
||||
release=worker.ReleaseRequest.from_task(task())
|
||||
calls=[]
|
||||
|
||||
def fake_http(base,method,route,token,body=None,timeout=30):
|
||||
calls.append((base,method,route,token,body))
|
||||
if route=="/health":
|
||||
return 200,{"status":"ok"}
|
||||
if route=="/createReleaseCandidate" and body=={}:
|
||||
return 400,{"error":"request_invalid"}
|
||||
if route.endswith("/claim"):
|
||||
return 200,{"lease_token":"l"*32,"attempt":1}
|
||||
if route.endswith("/heartbeat"):
|
||||
return 200,{}
|
||||
if route=="/createReleaseCandidate":
|
||||
return 200,{
|
||||
"candidate_id":release.candidate_id,
|
||||
"artifact_sha256":"f"*64,
|
||||
}
|
||||
if route.endswith("/result"):
|
||||
return 200,{"status":"VERIFYING"}
|
||||
raise AssertionError(route)
|
||||
|
||||
monkeypatch.setattr(worker,"http_json",fake_http)
|
||||
monkeypatch.setattr(worker,"audit",lambda *args,**kwargs:None)
|
||||
monkeypatch.setattr(worker,"write_liveness",lambda *args,**kwargs:None)
|
||||
monkeypatch.setattr(worker.Heartbeat,"first",lambda self:None)
|
||||
monkeypatch.setattr(worker.Heartbeat,"close",lambda self:None)
|
||||
monkeypatch.setattr(worker.Heartbeat,"ensure",lambda self:None)
|
||||
monkeypatch.setattr(worker.threading.Thread,"start",lambda self:None)
|
||||
|
||||
assert worker.execute_task(
|
||||
task(),
|
||||
agent_token="a"*32,
|
||||
release_token="r"*32,
|
||||
) is True
|
||||
|
||||
release_calls=[
|
||||
item for item in calls
|
||||
if item[0]==worker.RELEASE_API
|
||||
and item[2]=="/createReleaseCandidate"
|
||||
and item[4]!={}
|
||||
]
|
||||
assert len(release_calls)==1
|
||||
assert release_calls[0][1]=="POST"
|
||||
assert release_calls[0][2]=="/createReleaseCandidate"
|
||||
assert release_calls[0][4]=={
|
||||
"candidate_id":release.candidate_id,
|
||||
"source_commit":release.source_commit,
|
||||
"version":release.version,
|
||||
}
|
||||
|
||||
|
||||
|
||||
def test_write_liveness_records_release_capability(tmp_path,monkeypatch):
|
||||
path=tmp_path/"worker-liveness.json"
|
||||
monkeypatch.setattr(worker,"STATE_ROOT",tmp_path)
|
||||
monkeypatch.setattr(worker,"LIVENESS_PATH",path)
|
||||
worker.write_liveness(
|
||||
"IDLE",
|
||||
stage="poll",
|
||||
visible_count=2,
|
||||
)
|
||||
value=json.loads(path.read_text())
|
||||
assert value["schema"]=="aios.online-agent.worker-liveness.v1"
|
||||
assert value["agent_type"]=="release-preparation-worker"
|
||||
assert value["capabilities"]==["release:candidate"]
|
||||
assert value["state"]=="IDLE"
|
||||
assert value["visible_count"]==2
|
||||
|
||||
|
||||
def test_sd_notify_is_noop_without_systemd_socket(monkeypatch):
|
||||
monkeypatch.delenv("NOTIFY_SOCKET",raising=False)
|
||||
worker.sd_notify("WATCHDOG=1")
|
||||
|
||||
def test_systemd_unit_uses_two_encrypted_credentials_and_loopback_release_api():
|
||||
unit=(ROOT/"deploy/systemd/aios-release-online-agent-worker.service").read_text()
|
||||
assert "LoadCredentialEncrypted=agent_token:" in unit
|
||||
assert "LoadCredentialEncrypted=release_candidate_token:" in unit
|
||||
assert "AIOS_RELEASE_API_URL=http://127.0.0.1:18890" in unit
|
||||
assert "AIOS_RELEASE_WORKER_AGENT_ID=online-release-worker-cn-network-mkt-20260819" in unit
|
||||
assert "Type=notify" in unit
|
||||
assert "WatchdogSec=90" in unit
|
||||
assert "Restart=always" in unit
|
||||
assert "IPAddressDeny=any" in unit
|
||||
assert "IPAddressAllow=127.0.0.0/8" in unit
|
||||
assert "InaccessiblePaths=/data/aios-gitops /home" in unit
|
||||
assert "submitOwnerApproval" not in unit
|
||||
assert "startDeployment" not in unit
|
||||
|
||||
|
||||
def test_source_contains_no_high_risk_release_routes():
|
||||
source=(ROOT/"active-runtime/app/aios_release_online_agent_worker.py").read_text()
|
||||
assert '"/createReleaseCandidate"' in source
|
||||
for forbidden in (
|
||||
"/submitOwnerApproval",
|
||||
"/startDeployment",
|
||||
"/requestOrCreateTag",
|
||||
"/buildProductionArtifact",
|
||||
"/executeApprovedProductionRelease",
|
||||
):
|
||||
assert forbidden not in source
|
||||
|
||||
|
||||
def test_docs_keep_release_candidate_distinct_from_production():
|
||||
doc=(ROOT/"docs/superpowers/specs/2026-08-19-release-preparation-online-agent.md").read_text()
|
||||
assert "release:candidate" in doc
|
||||
assert "Owner signature" in doc
|
||||
assert "never\nstarts deployment" in doc or "never starts deployment" in doc
|
||||
assert "LoadCredentialEncrypted" in doc
|
||||
@@ -0,0 +1,116 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from pathlib import Path
|
||||
import json
|
||||
import sys
|
||||
|
||||
import pytest
|
||||
|
||||
|
||||
ROOT=Path(__file__).resolve().parents[1]
|
||||
sys.path.insert(0,str(ROOT/"active-runtime/app"))
|
||||
|
||||
import aios_release_preparation_supervisor_coordinator as coordinator
|
||||
|
||||
|
||||
def task():
|
||||
return {
|
||||
"task_id":coordinator.TASK_ID,
|
||||
"status":"QUEUED",
|
||||
"required_capabilities":coordinator.REQUIRED_CAPABILITIES,
|
||||
"metadata":{
|
||||
"execution":{
|
||||
"action":"agent_runtime",
|
||||
"adapter":"openclaw",
|
||||
"capability_scope":coordinator.REQUIRED_CAPABILITIES,
|
||||
"execution_digest":"e"*64,
|
||||
"change_set":[],
|
||||
"verification":[],
|
||||
"governance":{
|
||||
"writer_lease":False,
|
||||
"main_write":False,
|
||||
"production_write":False,
|
||||
"tag_write":False,
|
||||
"baseline_write":False,
|
||||
"force_push":False,
|
||||
"protected_ref_bypass":False,
|
||||
},
|
||||
}
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def expected_digest(monkeypatch):
|
||||
monkeypatch.setattr(coordinator,"EXPECTED_EXECUTION_DIGEST","e"*64)
|
||||
|
||||
|
||||
def test_coordinator_accepts_only_exact_nonmutating_projection():
|
||||
execution=coordinator.validate_task(task())
|
||||
assert execution["adapter"]=="openclaw"
|
||||
assert execution["governance"]["writer_lease"] is False
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"mutate,error",
|
||||
[
|
||||
(lambda value:value.__setitem__("required_capabilities",["agent.runtime.openclaw"]),
|
||||
"CAPABILITIES_INVALID"),
|
||||
(lambda value:value["metadata"]["execution"].__setitem__("execution_digest","f"*64),
|
||||
"ENVELOPE_INVALID"),
|
||||
(lambda value:value["metadata"]["execution"]["governance"].__setitem__("production_write",True),
|
||||
"ENVELOPE_INVALID"),
|
||||
],
|
||||
)
|
||||
def test_coordinator_rejects_capability_or_governance_drift(mutate,error):
|
||||
value=task();mutate(value)
|
||||
with pytest.raises(coordinator.CoordinatorError,match=error):
|
||||
coordinator.validate_task(value)
|
||||
|
||||
|
||||
def test_heartbeat_is_control_plane_host_only_and_never_forwards_credentials():
|
||||
progress=coordinator.heartbeat_progress(7)
|
||||
assert progress["control_plane_access_granted"] is True
|
||||
assert progress["external_planner_role"]=="UNPRIVILEGED_PLANNER"
|
||||
assert progress["credential_forwarded_to_planner"] is False
|
||||
assert progress["main_write"] is False
|
||||
assert progress["production_write"] is False
|
||||
encoded=json.dumps(progress,sort_keys=True).lower()
|
||||
assert "bearer " not in encoded
|
||||
assert "lease_token" not in encoded
|
||||
|
||||
|
||||
def test_coordinate_claim_heartbeats_without_submitting_result(monkeypatch,tmp_path):
|
||||
calls=[]
|
||||
def fake(method,route,token,body=None):
|
||||
calls.append((method,route,token,body))
|
||||
return 200,{"status":"RUNNING","lease_expires_at":"2099-01-01T00:00:00Z"}
|
||||
monkeypatch.setattr(coordinator,"api_call",fake)
|
||||
monkeypatch.setattr(coordinator,"STATE_ROOT",tmp_path)
|
||||
monkeypatch.setattr(coordinator,"HEARTBEAT_SECONDS",0)
|
||||
count=coordinator.coordinate_claim(
|
||||
{"lease_token":"l"*32},"agent-token-value-for-test",max_heartbeats=2
|
||||
)
|
||||
assert count==2
|
||||
assert all(call[1].endswith("/heartbeat") for call in calls)
|
||||
assert not any(call[1].endswith("/result") for call in calls)
|
||||
assert all("agent-token" not in json.dumps(call[3]) for call in calls)
|
||||
liveness=json.loads((tmp_path/"liveness.json").read_text())
|
||||
assert liveness["event_type"]=="SUPERVISOR_COORDINATOR_HEARTBEAT"
|
||||
assert liveness["credential_material_exposed"] is False
|
||||
|
||||
|
||||
def test_units_use_existing_encrypted_agent_credential_and_no_planner_credential():
|
||||
coordinator_unit=(
|
||||
ROOT/"deploy/systemd/aios-release-preparation-supervisor-coordinator.service"
|
||||
).read_text()
|
||||
projection_unit=(
|
||||
ROOT/"deploy/systemd/aios-agent-task-supervisor-coordinator-projection.service"
|
||||
).read_text()
|
||||
assert "aios-autonomous-agent-register.service" in coordinator_unit
|
||||
assert "LoadCredentialEncrypted=agent_token:" in coordinator_unit
|
||||
assert "control_token" not in coordinator_unit
|
||||
assert "SupplementaryGroups=aios-agent-runtime" in coordinator_unit
|
||||
assert "LoadCredential=agent_task_control_token:" in projection_unit
|
||||
assert "User=aios-agent-task" in projection_unit
|
||||
assert "ReadWritePaths=/var/lib/aios-agent-task" in projection_unit
|
||||
@@ -0,0 +1,215 @@
|
||||
import ast
|
||||
import json
|
||||
import os
|
||||
import struct
|
||||
from collections.abc import Mapping
|
||||
from pathlib import Path
|
||||
from types import SimpleNamespace
|
||||
from typing import Any
|
||||
|
||||
import pytest
|
||||
|
||||
from aios_agent_task_protocol import TaskProtocolError, TaskStore
|
||||
from aios_chatgpt_bridge_v2_server import BridgeHandler
|
||||
import aios_durable_owner_authorization as owner_auth
|
||||
from aios_online_agent_execution import GovernedExecutionBinding, project_online_task
|
||||
from aios_online_agent_worker import WorkerError, WorkerSupervisor
|
||||
|
||||
CAPABILITIES = ["git.read", "git.test", "governed.write.branch", "governed.write.file"]
|
||||
|
||||
|
||||
def binding(task_id="unified-test-task-20260824"):
|
||||
return GovernedExecutionBinding.from_mapping({
|
||||
"task_id": task_id,
|
||||
"job_id": "job-" + "a" * 32,
|
||||
"repository": "aios/aios",
|
||||
"base_oid": "0" * 40,
|
||||
"expected_branch_oid": "0" * 40,
|
||||
"target_branch": "agent/unified-test-20260824",
|
||||
"branch_mode": "CREATE_NEW",
|
||||
"generation": 1,
|
||||
"idempotency_key": "unified-test-task-20260824-v1",
|
||||
})
|
||||
|
||||
|
||||
def projected(task_id="unified-test-task-20260824"):
|
||||
return project_online_task(
|
||||
binding(task_id),
|
||||
{
|
||||
"goal": "test governed task",
|
||||
"instructions": "test only",
|
||||
"success_criteria": [],
|
||||
"required_capabilities": CAPABILITIES,
|
||||
"priority": 50,
|
||||
"max_attempts": 3,
|
||||
},
|
||||
{
|
||||
"action": "source_capture",
|
||||
"adapter": "source_capture",
|
||||
"capability_scope": CAPABILITIES,
|
||||
"source_signals": ["durable_policy"],
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
def test_write_task_rejected_before_persistence(tmp_path):
|
||||
store = TaskStore(tmp_path / "state.sqlite3")
|
||||
with pytest.raises(TaskProtocolError) as error:
|
||||
store.create_task(
|
||||
task_id="bare-write-task-20260824",
|
||||
goal="bare",
|
||||
instructions="must reject",
|
||||
success_criteria=[],
|
||||
required_capabilities=CAPABILITIES,
|
||||
metadata={},
|
||||
idempotency_key="bare-write-task-20260824-v1",
|
||||
)
|
||||
assert error.value.code == "GOVERNED_EXECUTION_BINDING_REQUIRED"
|
||||
with pytest.raises(TaskProtocolError) as missing:
|
||||
store.get_task("bare-write-task-20260824")
|
||||
assert missing.value.code == "TASK_NOT_FOUND"
|
||||
|
||||
|
||||
def test_complete_projection_persists_envelope(tmp_path):
|
||||
store = TaskStore(tmp_path / "state.sqlite3")
|
||||
task = projected()
|
||||
created = store.create_task(
|
||||
task_id=task["task_id"],
|
||||
goal=task["goal"],
|
||||
instructions=task["instructions"],
|
||||
success_criteria=task["success_criteria"],
|
||||
required_capabilities=task["required_capabilities"],
|
||||
constraints={},
|
||||
metadata=task["metadata"],
|
||||
priority=task["priority"],
|
||||
max_attempts=task["max_attempts"],
|
||||
idempotency_key=task["idempotency_key"],
|
||||
)
|
||||
execution = created["metadata"]["execution"]
|
||||
assert created["status"] == "QUEUED"
|
||||
assert execution["schema"] == "aios.online-agent.execution.v2"
|
||||
assert execution["job_id"].startswith("job-")
|
||||
assert execution["execution_digest"]
|
||||
assert created["metadata"]["execution_projection"]["execution_digest"] == execution["execution_digest"]
|
||||
|
||||
|
||||
def test_online_route_requires_governed_creation_for_write_caps():
|
||||
source = Path(__file__).parents[1] / "active-runtime" / "app" / "aios_model_api_staging.py"
|
||||
tree = ast.parse(source.read_text(encoding="utf-8"))
|
||||
function = next(node for node in tree.body if isinstance(node, ast.FunctionDef) and node.name == "requires_governed_execution")
|
||||
namespace = {
|
||||
"Mapping": Mapping,
|
||||
"Any": Any,
|
||||
"GOVERNED_EXECUTION_CAPABILITIES": {"governed.write.branch", "governed.write.file"},
|
||||
}
|
||||
exec(compile(ast.Module(body=[function], type_ignores=[]), str(source), "exec"), namespace)
|
||||
guard = namespace["requires_governed_execution"]
|
||||
assert guard({"required_capabilities": CAPABILITIES})
|
||||
assert not guard({"required_capabilities": ["git.read", "git.test"]})
|
||||
|
||||
|
||||
def test_bridge_requires_claim_for_projection_and_job(monkeypatch):
|
||||
handler = object.__new__(BridgeHandler)
|
||||
handler.headers = {"Authorization": "Bearer agent-credential"}
|
||||
handler.control_token = None
|
||||
handler.agent_task_api_url = "http://task-api"
|
||||
|
||||
class Response:
|
||||
def __init__(self, value):
|
||||
self.value = value
|
||||
def __enter__(self):
|
||||
return self
|
||||
def __exit__(self, *_args):
|
||||
return False
|
||||
def read(self):
|
||||
return json.dumps(self.value).encode()
|
||||
|
||||
responses = iter([
|
||||
{"agent_id": "worker-1"},
|
||||
{"task_id": "task-1", "claimed_by": None, "status": "QUEUED"},
|
||||
])
|
||||
monkeypatch.setattr("aios_chatgpt_bridge_v2_server.urllib.request.urlopen", lambda *_args, **_kwargs: Response(next(responses)))
|
||||
assert not handler._authorized("projectOnlineAgentExecution", {"task_id": "task-1"})
|
||||
|
||||
responses = iter([
|
||||
{"agent_id": "worker-1"},
|
||||
{"task_id": "task-1", "claimed_by": "worker-1", "status": "CLAIMED"},
|
||||
])
|
||||
assert handler._authorized("getGitOpsJob", {"task_id": "task-1"})
|
||||
|
||||
|
||||
def test_broker_uses_so_peercred_and_canonical_group(monkeypatch):
|
||||
class Connection:
|
||||
def getsockopt(self, *_args):
|
||||
return struct.pack("3i", os.getpid(), os.getuid(), os.getgid())
|
||||
|
||||
monkeypatch.setattr(owner_auth, "_peer_process_status", lambda _pid: (os.getuid(), os.getgid(), {os.getgid(), 4242}))
|
||||
monkeypatch.setattr(owner_auth, "CANONICAL_BROKER_GROUP", "aios-autonomous-agent-worker")
|
||||
monkeypatch.setattr(owner_auth.grp, "getgrnam", lambda _name: SimpleNamespace(gr_gid=4242))
|
||||
assert owner_auth.authorize_broker_peer(Connection())
|
||||
|
||||
monkeypatch.setattr(owner_auth, "_peer_process_status", lambda _pid: (os.getuid(), os.getgid(), {os.getgid()}))
|
||||
with pytest.raises(owner_auth.AuthorizationError) as error:
|
||||
owner_auth.authorize_broker_peer(Connection())
|
||||
assert error.value.code == "SOCKGROUP_NOT_AUTHORIZED"
|
||||
|
||||
|
||||
def test_bootstrap_rejects_bad_projection_before_job(tmp_path):
|
||||
from aios_governed_task_bootstrap import BootstrapError, GovernedTaskBootstrap
|
||||
|
||||
class Authority:
|
||||
def commit_exists(self, _oid):
|
||||
return True
|
||||
def read_ref(self, _ref):
|
||||
return None
|
||||
|
||||
bootstrap = GovernedTaskBootstrap(
|
||||
Authority(), tmp_path, remote_name="gitea-http",
|
||||
token_path=tmp_path / "token", ssh_key_path=tmp_path / "key",
|
||||
)
|
||||
with pytest.raises(BootstrapError) as error:
|
||||
bootstrap.create({
|
||||
"task_id": "atomic-projection-failure-20260824",
|
||||
"repository": "aios/aios",
|
||||
"base_oid": "0" * 40,
|
||||
"expected_branch_oid": "0" * 40,
|
||||
"target_branch": "agent/atomic-projection-failure-20260824",
|
||||
"risk": "engineering",
|
||||
"description": "invalid deterministic placeholder",
|
||||
"idempotency_key": "atomic-projection-failure-20260824-v1",
|
||||
"online_task": {
|
||||
"goal": "invalid",
|
||||
"instructions": "invalid",
|
||||
"success_criteria": [],
|
||||
"required_capabilities": CAPABILITIES,
|
||||
},
|
||||
"execution": {
|
||||
"action": "deterministic",
|
||||
"adapter": "deterministic",
|
||||
"capability_scope": CAPABILITIES,
|
||||
"change_set": [],
|
||||
"verification": [],
|
||||
},
|
||||
"require_online_task_projection": True,
|
||||
})
|
||||
assert error.value.args[0] == "EXECUTION_BINDING_INVALID"
|
||||
assert not list((tmp_path / "jobs").rglob("state.sqlite3"))
|
||||
assert not list((tmp_path / "task-bindings").glob("*.json"))
|
||||
|
||||
|
||||
def test_worker_rejects_bare_write_task():
|
||||
supervisor = object.__new__(WorkerSupervisor)
|
||||
with pytest.raises(WorkerError) as error:
|
||||
supervisor.resolve_envelope({
|
||||
"task_id": "bare-worker-task-20260824",
|
||||
"required_capabilities": CAPABILITIES,
|
||||
"metadata": {},
|
||||
})
|
||||
assert error.value.code == "TASK_SKIPPED_PROJECTION_MISSING"
|
||||
|
||||
|
||||
def test_continuation_scope_allows_existing_subset():
|
||||
source = Path(__file__).parents[1] / "active-runtime" / "app" / "aios_continuation_executor.py"
|
||||
text = source.read_text(encoding="utf-8")
|
||||
assert "if dirty and not dirty.issubset(expected_paths):" in text
|
||||
assert "if dirty and dirty != expected_paths:" not in text
|
||||
Reference in New Issue
Block a user