Compare commits
3
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
f71d678c2d | ||
|
|
f8e1022a41 | ||
|
|
2e5996463d |
@@ -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()
|
||||
@@ -0,0 +1,827 @@
|
||||
#!/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
|
||||
|
||||
|
||||
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"})
|
||||
|
||||
|
||||
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")
|
||||
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 = set(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 = set(json.loads(row["required_capabilities_json"]))
|
||||
if required.issubset(capabilities):
|
||||
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 = set(json.loads(row["required_capabilities_json"]))
|
||||
if not required.issubset(capabilities):
|
||||
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)
|
||||
]
|
||||
@@ -94,91 +94,6 @@ class AuthorityGitExecutor:
|
||||
self._git(["worktree", "add", "--detach", str(path), base_oid])
|
||||
return path
|
||||
|
||||
def recover_failed_detached_worktree_exact(
|
||||
self,
|
||||
*,
|
||||
path: Path,
|
||||
expected_oid: str,
|
||||
stale_oid: str,
|
||||
target_ref: str,
|
||||
) -> str:
|
||||
"""Recreate one exact clean, detached, single-attempt worktree at its durable head."""
|
||||
_oid(expected_oid, "expected_oid")
|
||||
_oid(stale_oid, "stale_oid")
|
||||
if not TASK_BRANCH_RE.fullmatch(target_ref):
|
||||
raise AuthorityGitError("TASK_BRANCH_SCOPE_DENIED")
|
||||
if self.read_ref(target_ref) is not None:
|
||||
raise AuthorityGitError("AUTHORITY_TARGET_BRANCH_ALREADY_EXISTS")
|
||||
if (
|
||||
not path.is_absolute()
|
||||
or path.parent != self.worktree_root
|
||||
or path.name in {"", ".", ".."}
|
||||
or path.is_symlink()
|
||||
or not path.is_dir()
|
||||
):
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_PATH_INVALID")
|
||||
head = subprocess.run(
|
||||
["git", "-C", str(path), "rev-parse", "HEAD"],
|
||||
text=True, capture_output=True, check=False,
|
||||
)
|
||||
if head.returncode != 0 or head.stdout.strip() != stale_oid or stale_oid == expected_oid:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_HEAD_INVALID")
|
||||
symbolic = subprocess.run(
|
||||
["git", "-C", str(path), "symbolic-ref", "-q", "HEAD"],
|
||||
text=True, capture_output=True, check=False,
|
||||
)
|
||||
if symbolic.returncode != 1:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_NOT_DETACHED")
|
||||
status = subprocess.run(
|
||||
["git", "-C", str(path), "status", "--porcelain=v1", "--untracked-files=all"],
|
||||
text=True, capture_output=True, check=False,
|
||||
)
|
||||
if status.returncode != 0 or status.stdout:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_NOT_CLEAN")
|
||||
parents = self._git(["rev-list", "--parents", "-n", "1", stale_oid]).split()
|
||||
if parents != [stale_oid, expected_oid]:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_LINEAGE_INVALID")
|
||||
registered = [
|
||||
line.removeprefix("worktree ")
|
||||
for line in self._git(["worktree", "list", "--porcelain"]).splitlines()
|
||||
if line.startswith("worktree ")
|
||||
]
|
||||
if registered.count(str(path)) != 1:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_REGISTRATION_INVALID")
|
||||
self._git(["worktree", "remove", "--force", str(path)])
|
||||
if path.exists() or path.is_symlink():
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_REMOVE_FAILED")
|
||||
recreated = self.create_detached_worktree(expected_oid, path.name, path.parent)
|
||||
readback = subprocess.run(
|
||||
["git", "-C", str(recreated), "rev-parse", "HEAD"],
|
||||
text=True, capture_output=True, check=False,
|
||||
)
|
||||
detached = subprocess.run(
|
||||
["git", "-C", str(recreated), "symbolic-ref", "-q", "HEAD"],
|
||||
text=True, capture_output=True, check=False,
|
||||
)
|
||||
if readback.returncode != 0 or readback.stdout.strip() != expected_oid or detached.returncode != 1:
|
||||
raise AuthorityGitError("WORKTREE_RECOVERY_READBACK_MISMATCH")
|
||||
return stale_oid
|
||||
|
||||
def create_task_branch_CAS(self, branch: str, expected_base_oid: str, new_oid: str) -> str:
|
||||
if not TASK_BRANCH_RE.fullmatch(branch):
|
||||
raise AuthorityGitError("TASK_BRANCH_SCOPE_DENIED")
|
||||
_oid(expected_base_oid, "expected_base_oid")
|
||||
_oid(new_oid, "new_oid")
|
||||
current = self.read_ref(branch)
|
||||
if current is not None:
|
||||
if current == new_oid:
|
||||
return new_oid
|
||||
raise AuthorityGitError("AUTHORITY_TARGET_BRANCH_ALREADY_EXISTS")
|
||||
if not self.ancestry_check(expected_base_oid, new_oid):
|
||||
raise AuthorityGitError("NEW_OID_NOT_DESCENDANT")
|
||||
self._git(["update-ref", branch, new_oid, "0" * 40])
|
||||
readback = self.read_ref(branch)
|
||||
if readback != new_oid:
|
||||
raise AuthorityGitError("AUTHORITY_CREATE_READBACK_MISMATCH")
|
||||
return readback
|
||||
|
||||
def advance_task_branch_CAS(self, branch: str, expected_oid: str, new_oid: str) -> str:
|
||||
if not TASK_BRANCH_RE.fullmatch(branch):
|
||||
raise AuthorityGitError("TASK_BRANCH_SCOPE_DENIED")
|
||||
|
||||
@@ -0,0 +1,133 @@
|
||||
"""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"
|
||||
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 != SCOPE:
|
||||
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)
|
||||
if value.get("effective_scopes") != [SCOPE]:
|
||||
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)
|
||||
@@ -13,8 +13,6 @@ BRIDGE_OPERATIONS = (
|
||||
"getGovernedWriteCapabilities",
|
||||
"createGovernedTask",
|
||||
"acquireWriterLease",
|
||||
"bootstrapWriterLease",
|
||||
"renewWriterLease",
|
||||
"commitFileGoverned",
|
||||
"commitFilesGoverned",
|
||||
"getGitOpsJob",
|
||||
@@ -33,8 +31,6 @@ BRIDGE_OPERATIONS = (
|
||||
OPERATION_REQUIRED_FIELDS = {
|
||||
"createGovernedTask": ("task_id", "repository", "base_oid", "target_branch", "risk", "description", "idempotency_key"),
|
||||
"acquireWriterLease": ("schema", "task_id", "repository", "branch_mode", "base_oid", "expected_branch_oid", "target_branch", "generation", "lease_seconds"),
|
||||
"bootstrapWriterLease": ("schema", "task_id", "repository", "branch_mode", "base_oid", "expected_branch_oid", "target_branch", "generation", "lease_seconds"),
|
||||
"renewWriterLease": ("schema", "task_id", "repository", "branch_mode", "base_oid", "expected_branch_oid", "target_branch", "generation", "lease_seconds"),
|
||||
"commitFileGoverned": ("task_id", "repository", "expected_branch_oid", "target_branch", "generation", "idempotency_key", "path", "content_base64", "message"),
|
||||
"commitFilesGoverned": ("task_id", "repository", "expected_branch_oid", "target_branch", "generation", "idempotency_key", "files", "message"),
|
||||
"getGitOpsJob": ("job_id",),
|
||||
@@ -50,49 +46,6 @@ OPERATION_REQUIRED_FIELDS = {
|
||||
"getControlPlaneSnapshot": (),
|
||||
}
|
||||
|
||||
BRIDGE_OPERATION_CONTRACTS = {
|
||||
"acquireWriterLease": {
|
||||
"method": "POST",
|
||||
"path": "/v2/writer-lease",
|
||||
"required_fields": OPERATION_REQUIRED_FIELDS["acquireWriterLease"],
|
||||
"response_fields": (
|
||||
"task_id", "job_id", "generation", "lease_seconds", "lease_expires_at",
|
||||
"create_target_branch", "worktree_mode", "cas_ref_update_required",
|
||||
),
|
||||
},
|
||||
"bootstrapWriterLease": {
|
||||
"method": "POST",
|
||||
"path": "/v2/writer-lease/bootstrap",
|
||||
"required_fields": OPERATION_REQUIRED_FIELDS["bootstrapWriterLease"],
|
||||
"response_fields": (
|
||||
"task_id", "job_id", "generation", "lease_seconds", "lease_expires_at",
|
||||
"create_target_branch", "worktree_mode", "cas_ref_update_required",
|
||||
"recovered", "bootstrapped", "runtime_rehydrated",
|
||||
"durable_branch_lease_present", "idempotent",
|
||||
),
|
||||
},
|
||||
"renewWriterLease": {
|
||||
"method": "POST",
|
||||
"path": "/v2/writer-lease/renew",
|
||||
"required_fields": OPERATION_REQUIRED_FIELDS["renewWriterLease"],
|
||||
"response_fields": (
|
||||
"task_id", "job_id", "generation", "lease_seconds", "lease_expires_at",
|
||||
"recovered", "bootstrapped", "runtime_rehydrated",
|
||||
"durable_branch_lease_present", "idempotent",
|
||||
),
|
||||
},
|
||||
"commitFileGoverned": {
|
||||
"method": "POST", "path": "/v2/commit-file",
|
||||
"required_fields": OPERATION_REQUIRED_FIELDS["commitFileGoverned"],
|
||||
"response_fields": ("job_id", "commit_oid", "commit_visible", "branch_mode", "target_branch", "generation"),
|
||||
},
|
||||
"commitFilesGoverned": {
|
||||
"method": "POST", "path": "/v2/commit-files",
|
||||
"required_fields": OPERATION_REQUIRED_FIELDS["commitFilesGoverned"],
|
||||
"response_fields": ("job_id", "commit_oid", "commit_visible", "branch_mode", "target_branch", "generation"),
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
class RuntimeDriftError(RuntimeError):
|
||||
pass
|
||||
@@ -169,14 +122,6 @@ class BridgeV2Adapter:
|
||||
"legacy_writes_enabled": False,
|
||||
"governed_writes_enabled": True,
|
||||
"operations": list(BRIDGE_OPERATIONS),
|
||||
"operation_contracts": {
|
||||
operation: {
|
||||
**contract,
|
||||
"required_fields": list(contract["required_fields"]),
|
||||
"response_fields": list(contract["response_fields"]),
|
||||
}
|
||||
for operation, contract in BRIDGE_OPERATION_CONTRACTS.items()
|
||||
},
|
||||
"force_push": False,
|
||||
"main_write": False,
|
||||
"tag_write": False,
|
||||
@@ -206,8 +151,6 @@ class BridgeV2Adapter:
|
||||
|
||||
def createGovernedTask(self, request: Mapping[str, Any]) -> Any: return self._call("createGovernedTask", request)
|
||||
def acquireWriterLease(self, request: Mapping[str, Any]) -> Any: return self._call("acquireWriterLease", request)
|
||||
def bootstrapWriterLease(self, request: Mapping[str, Any]) -> Any: return self._call("bootstrapWriterLease", request)
|
||||
def renewWriterLease(self, request: Mapping[str, Any]) -> Any: return self._call("renewWriterLease", request)
|
||||
def commitFileGoverned(self, request: Mapping[str, Any]) -> Any: return self._call("commitFileGoverned", request)
|
||||
def commitFilesGoverned(self, request: Mapping[str, Any]) -> Any: return self._call("commitFilesGoverned", request)
|
||||
def getGitOpsJob(self, request: Mapping[str, Any]) -> Any: return self._call("getGitOpsJob", request)
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@@ -51,13 +51,11 @@ class AuthorityPort(Protocol):
|
||||
def branch_exists(self, branch: str) -> bool: ...
|
||||
def branch_head(self, branch: str) -> str: ...
|
||||
def is_ancestor(self, older: str, newer: str) -> bool: ...
|
||||
def create_task_branch_cas(self, branch: str, expected_base: str, new: str) -> str: ...
|
||||
def compare_and_swap_branch(self, branch: str, expected: str, new: str) -> str: ...
|
||||
|
||||
|
||||
class RemotePort(Protocol):
|
||||
def repository_identity(self) -> str: ...
|
||||
def push_new_exact(self, branch: str, new: str) -> None: ...
|
||||
def push_exact(self, branch: str, expected: str, new: str) -> None: ...
|
||||
def read_git_ref(self, branch: str) -> str | None: ...
|
||||
def read_api_branch(self, branch: str) -> str | None: ...
|
||||
@@ -105,9 +103,6 @@ class AuthorityGitAdapter:
|
||||
def is_ancestor(self, older: str, newer: str) -> bool:
|
||||
return self.executor.ancestry_check(older, newer)
|
||||
|
||||
def create_task_branch_cas(self, branch: str, expected_base: str, new: str) -> str:
|
||||
return self.executor.create_task_branch_CAS(self._ref(branch), expected_base, new)
|
||||
|
||||
def compare_and_swap_branch(self, branch: str, expected: str, new: str) -> str:
|
||||
return self.executor.advance_task_branch_CAS(self._ref(branch), expected, new)
|
||||
|
||||
@@ -263,15 +258,6 @@ class GiteaContinuationRemote:
|
||||
raise ContinuationExecutionError("GITEA_GIT_READBACK_INVALID")
|
||||
return oid
|
||||
|
||||
def push_new_exact(self, branch: str, new: str) -> None:
|
||||
self._require_projected()
|
||||
self.repository_identity()
|
||||
if self.read_git_ref(branch) is not None:
|
||||
raise ContinuationExecutionError("GITEA_TARGET_BRANCH_ALREADY_EXISTS")
|
||||
_run(["git", "push", self.remote_name, f"{new}:refs/heads/{branch}"], cwd=self.workspace)
|
||||
if self.read_git_ref(branch) != new:
|
||||
raise ContinuationExecutionError("GITEA_CREATE_READBACK_MISMATCH")
|
||||
|
||||
def push_exact(self, branch: str, expected: str, new: str) -> None:
|
||||
self._require_projected()
|
||||
self.repository_identity()
|
||||
@@ -343,7 +329,6 @@ class ContinuationPermissions:
|
||||
@dataclass(frozen=True, slots=True)
|
||||
class ContinuationRuntimeBinding:
|
||||
request: WriterLeaseRequestV2
|
||||
job_id: str
|
||||
workspace_path: Path
|
||||
lease_not_before: datetime
|
||||
lease_expires_at: datetime
|
||||
@@ -404,20 +389,13 @@ class ContinuationExecutor:
|
||||
def observe_generation(self, branch: str, generation: int) -> None:
|
||||
self._observed_generations[branch] = max(generation, self._observed_generations.get(branch, 0))
|
||||
|
||||
def _max_generation(self, branch: str, *, exclude_job_id: str | None = None) -> int:
|
||||
def _max_generation(self, branch: str) -> int:
|
||||
observed = self._observed_generations.get(branch, 0)
|
||||
with self.controller.store._connect() as connection:
|
||||
if exclude_job_id is None:
|
||||
row = connection.execute(
|
||||
"SELECT MAX(generation) AS generation FROM branch_leases WHERE branch=?",
|
||||
(branch,),
|
||||
).fetchone()
|
||||
else:
|
||||
row = connection.execute(
|
||||
"SELECT MAX(generation) AS generation FROM branch_leases "
|
||||
"WHERE branch=? AND job_id!=?",
|
||||
(branch, exclude_job_id),
|
||||
).fetchone()
|
||||
durable = int(row["generation"] or 0)
|
||||
return max(observed, durable)
|
||||
|
||||
@@ -426,7 +404,6 @@ class ContinuationExecutor:
|
||||
return {
|
||||
"schema": WriterLeaseRequestV2.schema,
|
||||
"task_id": request.task_id,
|
||||
"job_id": request.job_id,
|
||||
"repository": request.repository,
|
||||
"branch_mode": request.branch_mode,
|
||||
"base_oid": request.base_oid,
|
||||
@@ -436,21 +413,10 @@ class ContinuationExecutor:
|
||||
"lease_seconds": request.lease_seconds,
|
||||
}
|
||||
|
||||
def _validate_preflight(self, binding: ContinuationRuntimeBinding, *, job_id: str | None = None) -> None:
|
||||
def _validate_preflight(self, binding: ContinuationRuntimeBinding) -> None:
|
||||
request = binding.request
|
||||
request_mapping = self._request_mapping(request)
|
||||
normalized = validate_writer_lease_request_v2(request_mapping, self.authority)
|
||||
expected_normalized = {
|
||||
**request_mapping,
|
||||
"create_target_branch": request.branch_mode == "CREATE_NEW",
|
||||
"worktree_mode": (
|
||||
"detached_at_base_oid"
|
||||
if request.branch_mode == "CREATE_NEW"
|
||||
else "detached_at_expected_head"
|
||||
),
|
||||
"cas_ref_update_required": True,
|
||||
}
|
||||
if normalized != expected_normalized:
|
||||
normalized = validate_writer_lease_request_v2(self._request_mapping(request), self.authority)
|
||||
if normalized != {**self._request_mapping(request), "create_target_branch": False, "worktree_mode": "detached_at_expected_head", "cas_ref_update_required": True}:
|
||||
raise ContinuationExecutionError("CONTINUATION_NORMALIZATION_INVALID")
|
||||
if not binding.original_receipt_valid or not binding.original_grant_consumed:
|
||||
raise ContinuationExecutionError("ORIGINAL_GRANT_LINEAGE_INVALID")
|
||||
@@ -460,7 +426,7 @@ class ContinuationExecutor:
|
||||
raise ContinuationExecutionError("LEASE_NOT_ACTIVE")
|
||||
if now >= binding.lease_expires_at:
|
||||
raise ContinuationExecutionError("LEASE_EXPIRED")
|
||||
if request.generation <= self._max_generation(request.target_branch, exclude_job_id=job_id):
|
||||
if request.generation <= self._max_generation(request.target_branch):
|
||||
raise ContinuationExecutionError("GENERATION_NOT_MONOTONIC")
|
||||
if not self.workspace.is_detached():
|
||||
raise ContinuationExecutionError("WORKTREE_NOT_DETACHED")
|
||||
@@ -486,102 +452,57 @@ class ContinuationExecutor:
|
||||
"generation": request.generation,
|
||||
}
|
||||
|
||||
def execute(
|
||||
self,
|
||||
binding: ContinuationRuntimeBinding,
|
||||
commit: ContinuationCommit,
|
||||
*,
|
||||
before_complete: Callable[[str], None] | None = None,
|
||||
) -> dict[str, object]:
|
||||
commit.payload()
|
||||
if not binding.job_id or binding.request.job_id != binding.job_id:
|
||||
raise ContinuationExecutionError("RUNTIME_LINEAGE_MISSING_JOB_ID")
|
||||
current = self.controller.store.read(binding.job_id)
|
||||
current_state = JobState(current["state"])
|
||||
if current_state == JobState.COMPLETED:
|
||||
def execute(self, binding: ContinuationRuntimeBinding, commit: ContinuationCommit) -> dict[str, object]:
|
||||
payload = commit.payload()
|
||||
submitted = self.controller.submit(self._job_request(binding, commit, payload), payload=payload)
|
||||
current = self.controller.store.read(submitted.job_id)
|
||||
if JobState(current["state"]) == JobState.COMPLETED:
|
||||
return self.controller.store._receipt(current, duplicate=True).as_dict()
|
||||
if current_state == JobState.REMOTE_VERIFIED:
|
||||
receipt = json.loads(str(current["receipt_json"] or "{}"))
|
||||
commit_oid = str(receipt.get("authority_oid") or "")
|
||||
if not OID_RE.fullmatch(commit_oid):
|
||||
raise ContinuationExecutionError("GOVERNED_COMMIT_OID_INVALID")
|
||||
if before_complete is not None:
|
||||
before_complete(commit_oid)
|
||||
return self.controller.advance(
|
||||
binding.job_id,
|
||||
JobState.COMPLETED,
|
||||
side_effect_state="REMOTE_VERIFIED",
|
||||
receipt=receipt,
|
||||
).as_dict()
|
||||
self._validate_preflight(binding, job_id=binding.job_id)
|
||||
self._validate_preflight(binding)
|
||||
request = binding.request
|
||||
projected = False
|
||||
try:
|
||||
self.credential.project()
|
||||
projected = True
|
||||
row = self.controller.store.read(binding.job_id)
|
||||
state = JobState(row["state"])
|
||||
if state == JobState.REQUESTED:
|
||||
row = self.controller.store.read(submitted.job_id)
|
||||
if JobState(row["state"]) == JobState.REQUESTED:
|
||||
self.controller.advance(
|
||||
binding.job_id,
|
||||
submitted.job_id,
|
||||
JobState.AUTHORIZED,
|
||||
side_effect_state="AUTHORIZED",
|
||||
receipt={"task_id": request.task_id, "original_grant_consumed": True},
|
||||
expected_revision=int(row["revision"]),
|
||||
)
|
||||
row = self.controller.store.read(binding.job_id)
|
||||
state = JobState(row["state"])
|
||||
if state == JobState.AUTHORIZED:
|
||||
row = self.controller.store.read(submitted.job_id)
|
||||
if JobState(row["state"]) == JobState.AUTHORIZED:
|
||||
self.controller.lease_job(
|
||||
binding.job_id,
|
||||
lease_id=f"lease-{binding.job_id}",
|
||||
submitted.job_id,
|
||||
lease_id=f"lease-{submitted.job_id}",
|
||||
lineage_anchor_oid=request.base_oid,
|
||||
current_head_oid=request.expected_branch_oid,
|
||||
generation=request.generation,
|
||||
lease_expires_at=binding.lease_expires_at.astimezone(timezone.utc).isoformat(),
|
||||
expected_revision=int(row["revision"]),
|
||||
)
|
||||
row = self.controller.store.read(binding.job_id)
|
||||
state = JobState(row["state"])
|
||||
if state == JobState.LEASED:
|
||||
self.controller.advance(binding.job_id, JobState.WORKTREE_READY,
|
||||
side_effect_state="WORKTREE_READY", expected_revision=int(row["revision"]))
|
||||
row = self.controller.store.read(binding.job_id)
|
||||
state = JobState(row["state"])
|
||||
if state == JobState.WORKTREE_READY:
|
||||
self.controller.advance(binding.job_id, JobState.EXECUTING,
|
||||
side_effect_state="IN_PROGRESS", expected_revision=int(row["revision"]))
|
||||
elif state != JobState.EXECUTING:
|
||||
raise ContinuationExecutionError("JOB_EXECUTION_STATE_INVALID")
|
||||
self.controller.advance(submitted.job_id, JobState.WORKTREE_READY, side_effect_state="WORKTREE_READY")
|
||||
self.controller.advance(submitted.job_id, JobState.EXECUTING, side_effect_state="IN_PROGRESS")
|
||||
new_oid = self.workspace.commit_files(commit.files, commit.message)
|
||||
if not OID_RE.fullmatch(new_oid) or not self.authority.is_ancestor(request.expected_branch_oid, new_oid):
|
||||
raise ContinuationExecutionError("NON_FAST_FORWARD_COMMIT")
|
||||
if request.branch_mode == "CREATE_NEW":
|
||||
authority_oid = self.authority.create_task_branch_cas(request.target_branch, request.base_oid, new_oid)
|
||||
else:
|
||||
authority_oid = self.authority.compare_and_swap_branch(request.target_branch, request.expected_branch_oid, new_oid)
|
||||
if authority_oid != new_oid:
|
||||
raise ContinuationExecutionError("AUTHORITY_CAS_READBACK_MISMATCH")
|
||||
self.controller.advance(
|
||||
binding.job_id,
|
||||
submitted.job_id,
|
||||
JobState.LOCAL_VERIFIED,
|
||||
side_effect_state="LOCAL_VERIFIED",
|
||||
receipt={"authority_oid": authority_oid, "force_push": False},
|
||||
)
|
||||
try:
|
||||
self.controller.advance(binding.job_id, JobState.REMOTE_SYNCING, side_effect_state="IN_PROGRESS")
|
||||
if request.branch_mode == "CREATE_NEW":
|
||||
self.remote.push_new_exact(request.target_branch, new_oid)
|
||||
else:
|
||||
self.controller.advance(submitted.job_id, JobState.REMOTE_SYNCING, side_effect_state="IN_PROGRESS")
|
||||
self.remote.push_exact(request.target_branch, request.expected_branch_oid, new_oid)
|
||||
except Exception:
|
||||
return self._remote_pending(binding.job_id, authority_oid, new_oid)
|
||||
return self._finish_remote_verification(
|
||||
binding.job_id,
|
||||
authority_oid,
|
||||
new_oid,
|
||||
before_complete=before_complete,
|
||||
)
|
||||
return self._remote_pending(submitted.job_id, authority_oid, new_oid)
|
||||
return self._finish_remote_verification(submitted.job_id, authority_oid, new_oid)
|
||||
finally:
|
||||
if projected:
|
||||
self.credential.revoke()
|
||||
@@ -604,14 +525,7 @@ class ContinuationExecutor:
|
||||
)
|
||||
return result.as_dict()
|
||||
|
||||
def _finish_remote_verification(
|
||||
self,
|
||||
job_id: str,
|
||||
authority_oid: str,
|
||||
new_oid: str,
|
||||
*,
|
||||
before_complete: Callable[[str], None] | None = None,
|
||||
) -> dict[str, object]:
|
||||
def _finish_remote_verification(self, job_id: str, authority_oid: str, new_oid: str) -> dict[str, object]:
|
||||
row = self.controller.store.read(job_id)
|
||||
branch = row["branch"]
|
||||
request = json.loads(row["request_json"])["binding"]
|
||||
@@ -628,20 +542,8 @@ class ContinuationExecutor:
|
||||
"reconciliation_decision": decision.action.value,
|
||||
"force_push": False,
|
||||
}
|
||||
self.controller.advance(
|
||||
job_id,
|
||||
JobState.REMOTE_VERIFIED,
|
||||
side_effect_state="REMOTE_VERIFIED",
|
||||
receipt=receipt,
|
||||
)
|
||||
if before_complete is not None:
|
||||
before_complete(new_oid)
|
||||
return self.controller.advance(
|
||||
job_id,
|
||||
JobState.COMPLETED,
|
||||
side_effect_state="REMOTE_VERIFIED",
|
||||
receipt=receipt,
|
||||
).as_dict()
|
||||
self.controller.advance(job_id, JobState.REMOTE_VERIFIED, side_effect_state="REMOTE_VERIFIED", receipt=receipt)
|
||||
return self.controller.advance(job_id, JobState.COMPLETED, side_effect_state="REMOTE_VERIFIED", receipt=receipt).as_dict()
|
||||
|
||||
def reconcile(self, job_id: str) -> dict[str, object]:
|
||||
row = self.controller.store.read(job_id)
|
||||
|
||||
@@ -49,7 +49,6 @@ class WriterLeaseRequestV2:
|
||||
target_branch: str
|
||||
generation: int
|
||||
lease_seconds: int = 900
|
||||
job_id: str = ""
|
||||
|
||||
schema = "aios.task-writer-lease.v2"
|
||||
accepted_schemas = frozenset({"aios.task-writer-lease.v2", "aios.task-writer-lease-request.v2"})
|
||||
@@ -57,8 +56,6 @@ class WriterLeaseRequestV2:
|
||||
def validate(self, probe: BranchProbe) -> None:
|
||||
if not self.task_id or self.repository != "aios/aios":
|
||||
raise ContinuationError("request_identity_invalid")
|
||||
if not self.job_id or not self.job_id.startswith("job-"):
|
||||
raise ContinuationError("job_id_invalid")
|
||||
if self.branch_mode not in {"CREATE_NEW", "CONTINUE_EXISTING"}:
|
||||
raise ContinuationError("branch_mode_invalid")
|
||||
_oid(self.base_oid, "base_oid")
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@@ -27,19 +27,6 @@ class ServiceUnavailable(JobError):
|
||||
pass
|
||||
|
||||
|
||||
@dataclass(frozen=True, slots=True)
|
||||
class LeaseRenewalResult:
|
||||
"""Durable outcome of one writer-lease recovery attempt."""
|
||||
|
||||
renewed: bool
|
||||
idempotent: bool
|
||||
durable: bool
|
||||
lease_expires_at: str | None
|
||||
|
||||
def __bool__(self) -> bool:
|
||||
return self.renewed or self.idempotent
|
||||
|
||||
|
||||
class BindingConflict(JobError):
|
||||
pass
|
||||
|
||||
@@ -79,14 +66,6 @@ def utc_now() -> str:
|
||||
return datetime.now(timezone.utc).isoformat().replace("+00:00", "Z")
|
||||
|
||||
|
||||
def parse_utc(value: str) -> datetime:
|
||||
normalized = value[:-1] + "+00:00" if value.endswith("Z") else value
|
||||
parsed = datetime.fromisoformat(normalized)
|
||||
if parsed.tzinfo is None:
|
||||
raise ValueError("LEASE_EXPIRY_INVALID")
|
||||
return parsed.astimezone(timezone.utc)
|
||||
|
||||
|
||||
def canonical_json(value: Any) -> bytes:
|
||||
return (
|
||||
json.dumps(value, sort_keys=True, separators=(",", ":"), ensure_ascii=False) + "\n"
|
||||
@@ -425,16 +404,13 @@ class GitOpsJobController:
|
||||
self.store._mirror(result)
|
||||
return self.store._receipt(result, duplicate=True)
|
||||
|
||||
def update_lease(self, lease_id: str, *, expected_revision: int, current_head_oid: str, generation: int, state: str = "ACTIVE", lease_expires_at: str | None = None) -> None:
|
||||
def update_lease(self, lease_id: str, *, expected_revision: int, current_head_oid: str, generation: int, state: str = "ACTIVE") -> None:
|
||||
if len(current_head_oid) != 40 or generation < 1:
|
||||
raise JobError("LEASE_BINDING_INVALID")
|
||||
try:
|
||||
with self.store._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
if lease_expires_at is None:
|
||||
updated = connection.execute("UPDATE branch_leases SET current_head_oid=?,generation=?,state=?,revision=revision+1 WHERE lease_id=? AND revision=?", (current_head_oid, generation, state, lease_id, expected_revision))
|
||||
else:
|
||||
updated = connection.execute("UPDATE branch_leases SET current_head_oid=?,generation=?,lease_expires_at=?,state=?,revision=revision+1 WHERE lease_id=? AND revision=?", (current_head_oid, generation, lease_expires_at, state, lease_id, expected_revision))
|
||||
if updated.rowcount != 1:
|
||||
raise JobError("LEASE_REVISION_CAS_CONFLICT")
|
||||
connection.commit()
|
||||
@@ -443,94 +419,6 @@ class GitOpsJobController:
|
||||
except sqlite3.Error as exc:
|
||||
raise ServiceUnavailable("JOB_DB_UNAVAILABLE") from exc
|
||||
|
||||
def renew_lease(
|
||||
self,
|
||||
*,
|
||||
branch: str,
|
||||
lineage_anchor_oid: str,
|
||||
current_head_oid: str,
|
||||
generation: int,
|
||||
lease_expires_at: str,
|
||||
) -> LeaseRenewalResult:
|
||||
if (
|
||||
not branch
|
||||
or len(lineage_anchor_oid) != 40
|
||||
or len(current_head_oid) != 40
|
||||
or generation < 1
|
||||
or not lease_expires_at.strip()
|
||||
):
|
||||
raise JobError("LEASE_BINDING_INVALID")
|
||||
try:
|
||||
with self.store._connect() as connection:
|
||||
connection.execute("BEGIN IMMEDIATE")
|
||||
rows = connection.execute(
|
||||
"SELECT lease_id, revision, generation, state, lease_expires_at "
|
||||
"FROM branch_leases "
|
||||
"WHERE branch=? AND lineage_anchor_oid=? AND current_head_oid=? "
|
||||
"AND generation=? AND state='ACTIVE' ORDER BY revision DESC",
|
||||
(branch, lineage_anchor_oid, current_head_oid, generation),
|
||||
).fetchall()
|
||||
if not rows:
|
||||
connection.commit()
|
||||
return LeaseRenewalResult(False, False, False, None)
|
||||
if len(rows) != 1:
|
||||
raise JobError("LEASE_BINDING_CONFLICT")
|
||||
row = rows[0]
|
||||
try:
|
||||
current_expires_at = parse_utc(str(row["lease_expires_at"]))
|
||||
requested_expires_at = parse_utc(lease_expires_at)
|
||||
except (TypeError, ValueError) as exc:
|
||||
raise JobError("LEASE_EXPIRY_INVALID") from exc
|
||||
now = datetime.now(timezone.utc)
|
||||
if requested_expires_at <= now:
|
||||
raise JobError("LEASE_EXPIRY_INVALID")
|
||||
if current_expires_at > now:
|
||||
connection.commit()
|
||||
return LeaseRenewalResult(
|
||||
False,
|
||||
True,
|
||||
True,
|
||||
str(row["lease_expires_at"]),
|
||||
)
|
||||
updated = connection.execute(
|
||||
"UPDATE branch_leases SET lease_expires_at=?,revision=revision+1 "
|
||||
"WHERE lease_id=? AND revision=? AND generation=? "
|
||||
"AND state='ACTIVE' AND lease_expires_at=?",
|
||||
(
|
||||
lease_expires_at,
|
||||
row["lease_id"],
|
||||
int(row["revision"]),
|
||||
generation,
|
||||
row["lease_expires_at"],
|
||||
),
|
||||
)
|
||||
if updated.rowcount != 1:
|
||||
current = connection.execute(
|
||||
"SELECT lease_expires_at, generation, state FROM branch_leases "
|
||||
"WHERE lease_id=?",
|
||||
(row["lease_id"],),
|
||||
).fetchone()
|
||||
if (
|
||||
current is not None
|
||||
and int(current["generation"]) == generation
|
||||
and current["state"] == "ACTIVE"
|
||||
and parse_utc(str(current["lease_expires_at"])) > now
|
||||
):
|
||||
connection.commit()
|
||||
return LeaseRenewalResult(
|
||||
False,
|
||||
True,
|
||||
True,
|
||||
str(current["lease_expires_at"]),
|
||||
)
|
||||
raise JobError("LEASE_REVISION_CAS_CONFLICT")
|
||||
connection.commit()
|
||||
return LeaseRenewalResult(True, False, True, lease_expires_at)
|
||||
except JobError:
|
||||
raise
|
||||
except sqlite3.Error as exc:
|
||||
raise ServiceUnavailable("JOB_DB_UNAVAILABLE") from exc
|
||||
|
||||
def advance(self, job_id: str, state: JobState, *, side_effect_state: str, receipt: Mapping[str, Any] | None = None, expected_revision: int | None = None) -> JobReceipt:
|
||||
return self._advance(
|
||||
job_id,
|
||||
|
||||
@@ -3,6 +3,13 @@ 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,
|
||||
CallerProjectionClient,
|
||||
CallerProjectionError,
|
||||
)
|
||||
from baseline_authority import BaselineAuthority, BaselineAuthorityError
|
||||
from typing import Any
|
||||
|
||||
@@ -89,6 +96,11 @@ 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 +124,9 @@ 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',
|
||||
@@ -407,14 +422,113 @@ 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','')
|
||||
try:
|
||||
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
|
||||
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 == SCOPE:
|
||||
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'],
|
||||
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 +560,183 @@ 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 +973,58 @@ 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]
|
||||
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':
|
||||
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 +1315,7 @@ 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 +1330,9 @@ 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,
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
@@ -964,7 +964,6 @@ def validate_writer_lease_request_v2(raw: Mapping[str, Any], probe: Any) -> dict
|
||||
required = {
|
||||
"schema",
|
||||
"task_id",
|
||||
"job_id",
|
||||
"repository",
|
||||
"branch_mode",
|
||||
"base_oid",
|
||||
@@ -977,7 +976,6 @@ def validate_writer_lease_request_v2(raw: Mapping[str, Any], probe: Any) -> dict
|
||||
raise ProvisioningError("LEASE_REQUEST_V2_FIELDS_REJECTED", 409)
|
||||
request = WriterLeaseRequestV2(
|
||||
task_id=str(raw["task_id"]),
|
||||
job_id=str(raw["job_id"]),
|
||||
repository=str(raw["repository"]),
|
||||
branch_mode=str(raw["branch_mode"]),
|
||||
base_oid=str(raw["base_oid"]),
|
||||
@@ -990,7 +988,6 @@ def validate_writer_lease_request_v2(raw: Mapping[str, Any], probe: Any) -> dict
|
||||
return {
|
||||
"schema": WriterLeaseRequestV2.schema,
|
||||
"task_id": request.task_id,
|
||||
"job_id": request.job_id,
|
||||
"repository": request.repository,
|
||||
"branch_mode": request.branch_mode,
|
||||
"base_oid": request.base_oid,
|
||||
|
||||
@@ -34,7 +34,6 @@ from .models import (
|
||||
)
|
||||
from .policy import PolicyEngine
|
||||
from .store import TaskPlaneStore
|
||||
from .scope import bind_task_scope
|
||||
from .worker_registry import WorkerRegistry
|
||||
|
||||
|
||||
@@ -69,54 +68,28 @@ class TaskPlaneService:
|
||||
self.delegation = DelegationEngine(store)
|
||||
self.store.probe_attestation_verifier = self.policy.probe_attestation_verifier
|
||||
|
||||
def _authorized_task(self, task_id: str, actor: str, action: str) -> TaskRecord:
|
||||
task = self.store.get_task(task_id, self.policy.actor_context(actor))
|
||||
self.policy.assert_allowed(action, actor, task=task)
|
||||
return task
|
||||
|
||||
def create_task(
|
||||
self,
|
||||
body: dict[str, Any],
|
||||
idempotency_key: str | None = None,
|
||||
*,
|
||||
actor_identity: str | None = None,
|
||||
) -> dict[str, Any]:
|
||||
def create_task(self, body: dict[str, Any], idempotency_key: str | None = None) -> dict[str, Any]:
|
||||
if body.get("parent_task_id"):
|
||||
raise ForbiddenError("child tasks must use the delegate endpoint")
|
||||
issuer = body.get("issuer", {})
|
||||
declared_actor = issuer.get("id", "anon") if isinstance(issuer, dict) else "anon"
|
||||
actor = actor_identity or declared_actor
|
||||
actor = issuer.get("id", "anon") if isinstance(issuer, dict) else "anon"
|
||||
self.policy.assert_allowed("create", actor)
|
||||
try:
|
||||
plan = TaskPlan.from_dict(body)
|
||||
except (TypeError, ValueError) as exc:
|
||||
raise ValidationError("invalid task plan") from exc
|
||||
if actor_identity is not None:
|
||||
plan = bind_task_scope(
|
||||
plan, self.policy.actor_context(actor_identity), public_create=True
|
||||
)
|
||||
task = self.store.create_task_with_event(plan, idempotency_key=idempotency_key)
|
||||
return task.to_dict()
|
||||
|
||||
def list_tasks(self, actor: str | None = None) -> list[dict[str, Any]]:
|
||||
context = None
|
||||
if actor is not None:
|
||||
self.policy.assert_allowed("read", actor)
|
||||
context = self.policy.actor_context(actor)
|
||||
return [
|
||||
task.to_dict(include_task=False)
|
||||
for task in self.store.list_tasks(context)
|
||||
]
|
||||
return [task.to_dict(include_task=False) for task in self.store.list_tasks()]
|
||||
|
||||
def get_task(self, task_id: str, actor: str | None = None) -> dict[str, Any]:
|
||||
context = None
|
||||
if actor is not None:
|
||||
self.policy.assert_allowed("read", actor)
|
||||
context = self.policy.actor_context(actor)
|
||||
task = self.store.get_task(task_id, context)
|
||||
if actor is not None:
|
||||
self.policy.assert_allowed("read", actor, task=task)
|
||||
return task.to_dict()
|
||||
self.policy.assert_allowed("read", actor, task=self.store.get_task(task_id))
|
||||
return self.store.get_task(task_id).to_dict()
|
||||
|
||||
def get_worker(self, worker_id: str, actor: str | None = None) -> dict[str, Any]:
|
||||
if actor is not None:
|
||||
@@ -190,8 +163,9 @@ class TaskPlaneService:
|
||||
if not worker_id:
|
||||
raise ValidationError("worker_id required")
|
||||
|
||||
task = self._authorized_task(task_id, actor, "claim")
|
||||
task = self.store.get_task(task_id)
|
||||
self.policy.assert_readonly_worker_binding(actor, str(worker_id))
|
||||
self.policy.assert_allowed("claim", actor, task=task)
|
||||
lease_seconds = _bounded_int(
|
||||
body.get("lease_seconds", 30), "lease_seconds", 1, 3600
|
||||
)
|
||||
@@ -210,9 +184,10 @@ class TaskPlaneService:
|
||||
if not claim_id:
|
||||
raise ValidationError("claim_id required")
|
||||
actor = body.get("actor_identity") or body.get("worker_identity") or "worker"
|
||||
task = self._authorized_task(task_id, actor, "heartbeat")
|
||||
task = self.store.get_task(task_id)
|
||||
claim = self.store.get_claim(claim_id)
|
||||
self.policy.assert_readonly_worker_binding(actor, claim.worker_id)
|
||||
self.policy.assert_allowed("heartbeat", actor, task=task)
|
||||
if claim.worker_identity != actor:
|
||||
raise ForbiddenError("claim identity mismatch")
|
||||
lease_seconds = _bounded_int(
|
||||
@@ -231,9 +206,10 @@ class TaskPlaneService:
|
||||
message = body.get("message")
|
||||
if progress is None:
|
||||
raise ValidationError("progress required")
|
||||
task = self._authorized_task(task_id, actor, "progress")
|
||||
task = self.store.get_task(task_id)
|
||||
claim = self.store.get_claim(claim_id)
|
||||
self.policy.assert_readonly_worker_binding(actor, claim.worker_id)
|
||||
self.policy.assert_allowed("progress", actor, task=task)
|
||||
if task.task_plan.data_classification == "credential-secret":
|
||||
if message not in {None, ""} or contains_secret_text(stage):
|
||||
raise ValidationError(
|
||||
@@ -258,7 +234,7 @@ class TaskPlaneService:
|
||||
if not claim_id:
|
||||
raise ValidationError("claim_id required")
|
||||
actor = body.get("actor_identity") or body.get("worker_identity") or "worker"
|
||||
task = self._authorized_task(task_id, actor, "result")
|
||||
task = self.store.get_task(task_id)
|
||||
|
||||
status = body.get("status", "COMPLETED")
|
||||
outputs = body.get("outputs", [])
|
||||
@@ -271,6 +247,7 @@ class TaskPlaneService:
|
||||
raise ValidationError("evidence items must be objects")
|
||||
claim = self.store.get_claim(claim_id)
|
||||
self.policy.assert_readonly_worker_binding(actor, claim.worker_id)
|
||||
self.policy.assert_allowed("result", actor, task=task)
|
||||
if claim.worker_identity != actor:
|
||||
raise ForbiddenError("claim identity mismatch")
|
||||
if task.task_plan.data_classification == "credential-secret":
|
||||
@@ -349,7 +326,8 @@ class TaskPlaneService:
|
||||
reason = body.get("reason", "failed")
|
||||
actor = body.get("actor_identity") or body.get("worker_identity") or "worker"
|
||||
self.policy.assert_allowed("fail", actor)
|
||||
task = self._authorized_task(task_id, actor, "fail")
|
||||
task = self.store.get_task(task_id)
|
||||
self.policy.assert_allowed("fail", actor, task=task)
|
||||
claim = self.store.get_claim(claim_id)
|
||||
if claim.worker_identity != actor:
|
||||
raise ForbiddenError("claim identity mismatch")
|
||||
@@ -361,7 +339,8 @@ class TaskPlaneService:
|
||||
actor = body.get("actor_identity") or body.get("worker_identity") or "owner"
|
||||
self.policy.assert_allowed("cancel", actor)
|
||||
reason = body.get("reason", "cancelled")
|
||||
task = self._authorized_task(task_id, actor, "cancel")
|
||||
task = self.store.get_task(task_id)
|
||||
self.policy.assert_allowed("cancel", actor, task=task)
|
||||
updated = self.store.cancel_task_with_event(
|
||||
task_id, task.revision, reason=reason
|
||||
)
|
||||
@@ -370,7 +349,8 @@ class TaskPlaneService:
|
||||
def delegate(self, task_id: str, body: dict[str, Any]) -> dict[str, Any]:
|
||||
actor = body.get("actor_identity") or "owner"
|
||||
self.policy.assert_allowed("delegate", actor)
|
||||
parent = self._authorized_task(task_id, actor, "delegate")
|
||||
parent = self.store.get_task(task_id)
|
||||
self.policy.assert_allowed("delegate", actor, task=parent)
|
||||
if "task" in body:
|
||||
try:
|
||||
child_plan = TaskPlan.from_dict(body["task"])
|
||||
@@ -384,30 +364,24 @@ class TaskPlaneService:
|
||||
except (TypeError, ValueError) as exc:
|
||||
raise ValidationError("invalid child task plan") from exc
|
||||
child_plan.parent_task_id = task_id
|
||||
child_plan = bind_task_scope(
|
||||
child_plan, self.policy.actor_context(actor), public_create=False
|
||||
)
|
||||
child = self.delegation.delegate(task_id, child_plan, caller=actor)
|
||||
return child.to_dict()
|
||||
|
||||
def get_events(self, task_id: str, actor: str | None = None) -> list[dict[str, Any]]:
|
||||
if actor is not None:
|
||||
self._authorized_task(task_id, actor, "audit")
|
||||
else:
|
||||
self.policy.assert_allowed("audit", actor, task=self.store.get_task(task_id))
|
||||
self.store.get_task(task_id)
|
||||
return self.store.list_events(task_id)
|
||||
|
||||
def get_evidence(self, task_id: str, actor: str | None = None) -> list[dict[str, Any]]:
|
||||
if actor is not None:
|
||||
self._authorized_task(task_id, actor, "audit")
|
||||
else:
|
||||
self.policy.assert_allowed("audit", actor, task=self.store.get_task(task_id))
|
||||
self.store.get_task(task_id)
|
||||
return self.evidence.list(task_id)
|
||||
|
||||
def get_children(self, task_id: str, actor: str | None = None) -> list[str]:
|
||||
if actor is not None:
|
||||
self._authorized_task(task_id, actor, "read")
|
||||
else:
|
||||
self.policy.assert_allowed("read", actor, task=self.store.get_task(task_id))
|
||||
self.store.get_task(task_id)
|
||||
return self.store.list_children(task_id)
|
||||
|
||||
@@ -532,11 +506,7 @@ class TaskPlaneRequestHandler(BaseHTTPRequestHandler):
|
||||
):
|
||||
raise ForbiddenError("issuer identity mismatch")
|
||||
body["issuer"] = {"type": "authenticated", "id": identity}
|
||||
return 201, self.service.create_task(
|
||||
body or {},
|
||||
headers.get("idempotency-key"),
|
||||
actor_identity=identity,
|
||||
)
|
||||
return 201, self.service.create_task(body or {}, headers.get("idempotency-key"))
|
||||
if body.get("actor_identity") not in {None, identity}:
|
||||
raise ForbiddenError("actor identity mismatch")
|
||||
body["actor_identity"] = identity
|
||||
|
||||
@@ -13,8 +13,6 @@ from dataclasses import dataclass, field
|
||||
from datetime import datetime, timezone
|
||||
from typing import Any
|
||||
|
||||
from .scope import normalize_task_scope
|
||||
|
||||
|
||||
SENSITIVE_KEYS = {
|
||||
"secret",
|
||||
@@ -68,12 +66,6 @@ TASK_PLAN_KEYS = {
|
||||
"allowed_input_paths",
|
||||
"allowed_output_paths",
|
||||
"output_policy",
|
||||
"scope_type",
|
||||
"tenant_id",
|
||||
"user_id",
|
||||
"project_id",
|
||||
"conversation_id",
|
||||
"run_id",
|
||||
}
|
||||
|
||||
|
||||
@@ -245,12 +237,6 @@ class TaskPlan:
|
||||
schema_version: int = 1
|
||||
task_id: str | None = None
|
||||
task_type: str = "code.change"
|
||||
scope_type: str = "legacy"
|
||||
tenant_id: str | None = None
|
||||
user_id: str | None = None
|
||||
project_id: str | None = None
|
||||
conversation_id: str | None = None
|
||||
run_id: str | None = None
|
||||
issuer: dict[str, Any] = field(default_factory=dict)
|
||||
intent: str | None = None
|
||||
required_capabilities: list[str] = field(default_factory=list)
|
||||
@@ -325,17 +311,10 @@ class TaskPlan:
|
||||
raise ValueError("credential-secret requires metadata-only logging")
|
||||
if output_policy != "hash-count-status-only":
|
||||
raise ValueError("credential-secret requires restricted output")
|
||||
scope = normalize_task_scope(data)
|
||||
return cls(
|
||||
schema_version=schema_version,
|
||||
task_id=validate_identifier(data["task_id"], "task_id"),
|
||||
task_type=validate_type(data.get("task_type", "code.change"), "task_type"),
|
||||
scope_type=scope.scope_type,
|
||||
tenant_id=scope.tenant_id,
|
||||
user_id=scope.user_id,
|
||||
project_id=scope.project_id,
|
||||
conversation_id=scope.conversation_id,
|
||||
run_id=scope.run_id,
|
||||
issuer=issuer,
|
||||
intent=data.get("intent"),
|
||||
required_capabilities=_normalize_list(data.get("required_capabilities", [])),
|
||||
@@ -369,12 +348,6 @@ class TaskPlan:
|
||||
"schema_version": self.schema_version,
|
||||
"task_id": self.task_id,
|
||||
"task_type": self.task_type,
|
||||
"scope_type": self.scope_type,
|
||||
"tenant_id": self.tenant_id,
|
||||
"user_id": self.user_id,
|
||||
"project_id": self.project_id,
|
||||
"conversation_id": self.conversation_id,
|
||||
"run_id": self.run_id,
|
||||
"issuer": self.issuer,
|
||||
"intent": self.intent,
|
||||
"required_capabilities": list(self.required_capabilities),
|
||||
|
||||
@@ -3,11 +3,10 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
from typing import Any, Callable, Mapping
|
||||
from typing import Callable
|
||||
|
||||
from .errors import ForbiddenError, LockedError
|
||||
from .models import ProbeEvidence, ResultEnvelope, TaskRecord
|
||||
from .scope import TaskActorContext, actor_context as build_actor_context
|
||||
|
||||
|
||||
RISK_RANK = {"low": 1, "medium": 2, "high": 3}
|
||||
@@ -94,10 +93,6 @@ class PolicyEngine:
|
||||
result_signature_verifier: Callable[[ResultEnvelope], bool] | None = None,
|
||||
probe_attestation_verifier: Callable[[ProbeEvidence], bool] | None = None,
|
||||
identity_capabilities: dict[str, set[str]] | None = None,
|
||||
identity_actor_contexts: Mapping[
|
||||
str, TaskActorContext | Mapping[str, Any]
|
||||
]
|
||||
| None = None,
|
||||
admin_identities: set[str] | None = None,
|
||||
readonly_worker_execution_enabled: bool = False,
|
||||
readonly_worker_profiles: list[ReadonlyWorkerProfile] | None = None,
|
||||
@@ -114,26 +109,12 @@ class PolicyEngine:
|
||||
identity: set(capabilities)
|
||||
for identity, capabilities in (identity_capabilities or {}).items()
|
||||
}
|
||||
self.identity_actor_contexts = dict(identity_actor_contexts or {})
|
||||
self.admin_identities = set(admin_identities or ())
|
||||
self.readonly_worker_execution_enabled = readonly_worker_execution_enabled
|
||||
self.readonly_worker_profiles = {
|
||||
profile.identity: profile for profile in (readonly_worker_profiles or [])
|
||||
}
|
||||
|
||||
def actor_context(self, identity: str) -> TaskActorContext:
|
||||
configured = self.identity_actor_contexts.get(
|
||||
identity, self.identity_actor_contexts.get("*")
|
||||
)
|
||||
capabilities = set(
|
||||
self.identity_capabilities.get(
|
||||
identity, self.identity_capabilities.get("*", set())
|
||||
)
|
||||
)
|
||||
if identity in self.admin_identities:
|
||||
capabilities.add("task.admin")
|
||||
return build_actor_context(identity, configured, capabilities)
|
||||
|
||||
def _readonly_worker_allows(
|
||||
self, action: str, actor: str, task: TaskRecord | None
|
||||
) -> bool:
|
||||
|
||||
@@ -1,181 +0,0 @@
|
||||
"""Trusted TaskPlane scope model and SQL authorization helpers."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass, replace
|
||||
from typing import Any, Mapping
|
||||
|
||||
from .errors import ForbiddenError, ValidationError
|
||||
|
||||
SCOPE_FIELDS = ("tenant_id", "user_id", "project_id", "conversation_id", "run_id")
|
||||
PRIVILEGED_CAPABILITIES = frozenset(
|
||||
{"*", "admin", "system.admin", "task.admin", "task.read.all"}
|
||||
)
|
||||
|
||||
|
||||
def _value(value: Any, field: str) -> Any:
|
||||
if hasattr(value, "task_plan"):
|
||||
value = value.task_plan
|
||||
if isinstance(value, Mapping):
|
||||
return value.get(field)
|
||||
return getattr(value, field, None)
|
||||
|
||||
|
||||
def _text(value: Any) -> str | None:
|
||||
if value is None:
|
||||
return None
|
||||
result = str(value).strip()
|
||||
return result or None
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class TaskScope:
|
||||
scope_type: str
|
||||
tenant_id: str | None = None
|
||||
user_id: str | None = None
|
||||
project_id: str | None = None
|
||||
conversation_id: str | None = None
|
||||
run_id: str | None = None
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class TaskActorContext:
|
||||
identity: str
|
||||
scope_type: str = "legacy"
|
||||
tenant_id: str | None = None
|
||||
user_id: str | None = None
|
||||
project_id: str | None = None
|
||||
conversation_id: str | None = None
|
||||
run_id: str | None = None
|
||||
capabilities: frozenset[str] = frozenset()
|
||||
|
||||
|
||||
def normalize_task_scope(value: Any) -> TaskScope:
|
||||
fields = {field: _text(_value(value, field)) for field in SCOPE_FIELDS}
|
||||
declared = _text(_value(value, "scope_type"))
|
||||
if declared is None:
|
||||
declared = "tenant" if fields["tenant_id"] else "legacy"
|
||||
if declared == "tenant":
|
||||
if fields["tenant_id"] is None:
|
||||
raise ValidationError(
|
||||
"tenant scope requires tenant_id", code="TASK_SCOPE_TENANT_ID_REQUIRED"
|
||||
)
|
||||
elif declared == "legacy":
|
||||
if any(fields.values()):
|
||||
raise ValidationError(
|
||||
"legacy scope cannot contain scoped ownership",
|
||||
code="TASK_SCOPE_LEGACY_WITH_SCOPED_FIELDS",
|
||||
)
|
||||
elif declared == "system":
|
||||
if any(fields[field] for field in ("tenant_id", "user_id", "project_id", "conversation_id")):
|
||||
raise ValidationError(
|
||||
"system scope cannot contain tenant ownership",
|
||||
code="TASK_SCOPE_SYSTEM_WITH_TENANT_FIELDS",
|
||||
)
|
||||
else:
|
||||
raise ValidationError("invalid task scope type", code="TASK_SCOPE_TYPE_INVALID")
|
||||
return TaskScope(declared, **fields)
|
||||
|
||||
|
||||
def actor_is_privileged(actor: TaskActorContext) -> bool:
|
||||
return bool(actor.capabilities & PRIVILEGED_CAPABILITIES)
|
||||
|
||||
|
||||
def actor_context(
|
||||
identity: str,
|
||||
configured: TaskActorContext | Mapping[str, Any] | None,
|
||||
capabilities: set[str] | frozenset[str],
|
||||
) -> TaskActorContext:
|
||||
if isinstance(configured, TaskActorContext):
|
||||
context = configured
|
||||
else:
|
||||
data = dict(configured or {})
|
||||
data.pop("identity", None)
|
||||
data.pop("capabilities", None)
|
||||
context = TaskActorContext(identity=str(identity), **data)
|
||||
scope = normalize_task_scope(context)
|
||||
return replace(
|
||||
context,
|
||||
identity=str(identity),
|
||||
scope_type=scope.scope_type,
|
||||
tenant_id=scope.tenant_id,
|
||||
user_id=scope.user_id,
|
||||
project_id=scope.project_id,
|
||||
conversation_id=scope.conversation_id,
|
||||
run_id=scope.run_id,
|
||||
capabilities=frozenset(str(item) for item in capabilities),
|
||||
)
|
||||
|
||||
|
||||
def bind_task_scope(plan: Any, actor: TaskActorContext, *, public_create: bool) -> Any:
|
||||
requested = normalize_task_scope(plan)
|
||||
if actor_is_privileged(actor):
|
||||
trusted = normalize_task_scope(actor)
|
||||
if requested.scope_type == "legacy" and trusted.scope_type != "legacy":
|
||||
plan.scope_type = trusted.scope_type
|
||||
for field in SCOPE_FIELDS:
|
||||
setattr(plan, field, getattr(actor, field))
|
||||
return plan
|
||||
trusted = normalize_task_scope(actor)
|
||||
if public_create and trusted.scope_type == "legacy":
|
||||
raise ForbiddenError(
|
||||
"authenticated identity has no tenant scope",
|
||||
code="TASK_ACTOR_SCOPE_UNAVAILABLE",
|
||||
)
|
||||
if trusted.scope_type != "tenant":
|
||||
if requested.scope_type == "legacy":
|
||||
plan.scope_type = trusted.scope_type
|
||||
for field in SCOPE_FIELDS:
|
||||
setattr(plan, field, getattr(actor, field))
|
||||
return plan
|
||||
if requested.scope_type != trusted.scope_type:
|
||||
raise ForbiddenError("task scope mismatch", code="TASK_SCOPE_MISMATCH_DENY")
|
||||
return plan
|
||||
if requested.scope_type not in {"legacy", "tenant"}:
|
||||
raise ForbiddenError("task scope mismatch", code="TASK_SCOPE_MISMATCH_DENY")
|
||||
updates: dict[str, Any] = {"scope_type": "tenant"}
|
||||
for field in SCOPE_FIELDS:
|
||||
declared = getattr(requested, field)
|
||||
authoritative = getattr(actor, field)
|
||||
if declared is not None and declared != authoritative:
|
||||
raise ForbiddenError(
|
||||
"task scope mismatch",
|
||||
code="TASK_SCOPE_MISMATCH_DENY",
|
||||
details={"field": field},
|
||||
)
|
||||
updates[field] = authoritative
|
||||
for field, item in updates.items():
|
||||
setattr(plan, field, item)
|
||||
return plan
|
||||
|
||||
|
||||
def actor_scope_sql(actor: TaskActorContext | None) -> tuple[str, tuple[Any, ...]]:
|
||||
if actor is None or actor_is_privileged(actor):
|
||||
return "", ()
|
||||
scope = normalize_task_scope(actor)
|
||||
if scope.scope_type == "legacy":
|
||||
return "scope_type = 'legacy'", ()
|
||||
if scope.scope_type == "system":
|
||||
return "scope_type = 'system'", ()
|
||||
clauses = ["scope_type = 'tenant'", "tenant_id = ?"]
|
||||
params: list[Any] = [scope.tenant_id]
|
||||
for field in ("user_id", "project_id", "conversation_id", "run_id"):
|
||||
value = getattr(scope, field)
|
||||
if value is None:
|
||||
clauses.append(f"{field} IS NULL")
|
||||
else:
|
||||
clauses.append(f"({field} IS NULL OR {field} = ?)")
|
||||
params.append(value)
|
||||
return " AND ".join(clauses), tuple(params)
|
||||
|
||||
|
||||
def task_scope_values(value: Any) -> tuple[str | None, ...]:
|
||||
scope = normalize_task_scope(value)
|
||||
return (
|
||||
scope.scope_type,
|
||||
scope.tenant_id,
|
||||
scope.user_id,
|
||||
scope.project_id,
|
||||
scope.conversation_id,
|
||||
scope.run_id,
|
||||
)
|
||||
@@ -32,7 +32,6 @@ from .models import (
|
||||
redact_value,
|
||||
validate_identifier,
|
||||
)
|
||||
from .scope import TaskActorContext, actor_scope_sql, task_scope_values
|
||||
|
||||
|
||||
EventEnvelope = tuple[str, str, str, str, str, float]
|
||||
@@ -135,12 +134,6 @@ class TaskPlaneStore:
|
||||
status TEXT NOT NULL,
|
||||
task_json TEXT NOT NULL,
|
||||
task_type TEXT NOT NULL,
|
||||
scope_type TEXT NOT NULL DEFAULT 'legacy',
|
||||
tenant_id TEXT,
|
||||
user_id TEXT,
|
||||
project_id TEXT,
|
||||
conversation_id TEXT,
|
||||
run_id TEXT,
|
||||
root_task_id TEXT NOT NULL,
|
||||
parent_task_id TEXT,
|
||||
correlation_id TEXT,
|
||||
@@ -279,21 +272,6 @@ class TaskPlaneStore:
|
||||
)
|
||||
"""
|
||||
)
|
||||
self._ensure_columns(
|
||||
"tasks",
|
||||
{
|
||||
"scope_type": "TEXT NOT NULL DEFAULT 'legacy'",
|
||||
"tenant_id": "TEXT",
|
||||
"user_id": "TEXT",
|
||||
"project_id": "TEXT",
|
||||
"conversation_id": "TEXT",
|
||||
"run_id": "TEXT",
|
||||
},
|
||||
)
|
||||
self.conn.execute(
|
||||
"UPDATE tasks SET scope_type='legacy' "
|
||||
"WHERE scope_type IS NULL OR TRIM(scope_type)=''"
|
||||
)
|
||||
self._ensure_columns(
|
||||
"workers",
|
||||
{
|
||||
@@ -321,14 +299,6 @@ class TaskPlaneStore:
|
||||
)
|
||||
self.conn.execute("CREATE INDEX IF NOT EXISTS ix_tasks_status ON tasks(status)")
|
||||
self.conn.execute("CREATE INDEX IF NOT EXISTS ix_tasks_parent ON tasks(parent_task_id)")
|
||||
self.conn.execute(
|
||||
"CREATE INDEX IF NOT EXISTS ix_tasks_scope_tenant_project "
|
||||
"ON tasks(scope_type, tenant_id, project_id)"
|
||||
)
|
||||
self.conn.execute(
|
||||
"CREATE INDEX IF NOT EXISTS ix_tasks_scope_owner "
|
||||
"ON tasks(scope_type, tenant_id, user_id, conversation_id, run_id)"
|
||||
)
|
||||
self.conn.execute("CREATE INDEX IF NOT EXISTS ix_workers_status ON workers(status)")
|
||||
self.conn.execute("CREATE INDEX IF NOT EXISTS ix_claims_task ON claims(task_id)")
|
||||
self.conn.execute(
|
||||
@@ -377,42 +347,6 @@ class TaskPlaneStore:
|
||||
raise NotFoundError("task")
|
||||
task_json = json.loads(row["task_json"])
|
||||
task_plan = TaskPlan.from_dict(task_json)
|
||||
stored_scope = tuple(
|
||||
row[field]
|
||||
for field in (
|
||||
"scope_type",
|
||||
"tenant_id",
|
||||
"user_id",
|
||||
"project_id",
|
||||
"conversation_id",
|
||||
"run_id",
|
||||
)
|
||||
)
|
||||
json_has_scope = any(
|
||||
field in task_json
|
||||
for field in (
|
||||
"scope_type",
|
||||
"tenant_id",
|
||||
"user_id",
|
||||
"project_id",
|
||||
"conversation_id",
|
||||
"run_id",
|
||||
)
|
||||
)
|
||||
if json_has_scope and stored_scope != task_scope_values(task_plan):
|
||||
raise ConflictError(
|
||||
"task scope persistence mismatch",
|
||||
code="TASK_SCOPE_PERSISTENCE_MISMATCH",
|
||||
)
|
||||
if not json_has_scope:
|
||||
(
|
||||
task_plan.scope_type,
|
||||
task_plan.tenant_id,
|
||||
task_plan.user_id,
|
||||
task_plan.project_id,
|
||||
task_plan.conversation_id,
|
||||
task_plan.run_id,
|
||||
) = stored_scope
|
||||
task_plan.parent_task_id = row["parent_task_id"]
|
||||
task_plan.correlation_id = row["correlation_id"]
|
||||
return TaskRecord(
|
||||
@@ -455,7 +389,6 @@ class TaskPlaneStore:
|
||||
task_id = task_plan.task_id or f"task-{uuid.uuid4()}"
|
||||
task_plan.task_id = task_id
|
||||
task_json = redact_value(task_plan.to_dict(redact=False))
|
||||
scope_values = task_scope_values(task_plan)
|
||||
now = _now()
|
||||
with self._transaction():
|
||||
if idempotency_key:
|
||||
@@ -464,11 +397,7 @@ class TaskPlaneStore:
|
||||
(idempotency_key,),
|
||||
).fetchone()
|
||||
if existing:
|
||||
existing_plan = TaskPlan.from_dict(
|
||||
json.loads(existing["task_json"])
|
||||
)
|
||||
existing_json = redact_value(existing_plan.to_dict(redact=False))
|
||||
if _canonical(existing_json) != _canonical(task_json):
|
||||
if _canonical(json.loads(existing["task_json"])) != _canonical(task_json):
|
||||
raise ConflictError("idempotency key payload mismatch")
|
||||
return self.get_task(existing["task_id"])
|
||||
root_task_id = task_id
|
||||
@@ -483,18 +412,14 @@ class TaskPlaneStore:
|
||||
self.conn.execute(
|
||||
"""
|
||||
INSERT INTO tasks
|
||||
(task_id, revision, status, task_json, task_type,
|
||||
scope_type, tenant_id, user_id, project_id, conversation_id, run_id,
|
||||
root_task_id, parent_task_id, correlation_id, idempotency_key,
|
||||
created_at, updated_at, failure_reason)
|
||||
VALUES (?, 1, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, NULL)
|
||||
(task_id, revision, status, task_json, task_type, root_task_id, parent_task_id, correlation_id, idempotency_key, created_at, updated_at, failure_reason)
|
||||
VALUES (?, 1, ?, ?, ?, ?, ?, ?, ?, ?, ?, NULL)
|
||||
""",
|
||||
(
|
||||
task_id,
|
||||
"CREATED",
|
||||
json.dumps(task_json, ensure_ascii=False),
|
||||
task_plan.task_type,
|
||||
*scope_values,
|
||||
root_task_id,
|
||||
task_plan.parent_task_id,
|
||||
task_plan.correlation_id,
|
||||
@@ -526,30 +451,21 @@ class TaskPlaneStore:
|
||||
)
|
||||
return self.get_task(task.task_id)
|
||||
|
||||
def get_task(
|
||||
self, task_id: str, actor_context: TaskActorContext | None = None
|
||||
) -> TaskRecord:
|
||||
def get_task(self, task_id: str) -> TaskRecord:
|
||||
task_id = _validated_id(task_id, "task_id")
|
||||
predicate, scope_params = actor_scope_sql(actor_context)
|
||||
sql = "SELECT * FROM tasks WHERE task_id=?"
|
||||
if predicate:
|
||||
sql += f" AND ({predicate})"
|
||||
with self._lock:
|
||||
row = self.conn.execute(sql, (task_id, *scope_params)).fetchone()
|
||||
row = self.conn.execute(
|
||||
"SELECT * FROM tasks WHERE task_id=?", (task_id,)
|
||||
).fetchone()
|
||||
if row is None:
|
||||
raise NotFoundError(f"task {task_id} not found")
|
||||
return self._row_to_task(row)
|
||||
|
||||
def list_tasks(
|
||||
self, actor_context: TaskActorContext | None = None
|
||||
) -> list[TaskRecord]:
|
||||
predicate, scope_params = actor_scope_sql(actor_context)
|
||||
sql = "SELECT * FROM tasks"
|
||||
if predicate:
|
||||
sql += f" WHERE ({predicate})"
|
||||
sql += " ORDER BY created_at DESC"
|
||||
def list_tasks(self) -> list[TaskRecord]:
|
||||
with self._lock:
|
||||
rows = self.conn.execute(sql, scope_params).fetchall()
|
||||
rows = self.conn.execute(
|
||||
"SELECT * FROM tasks ORDER BY created_at DESC"
|
||||
).fetchall()
|
||||
return [self._row_to_task(row) for row in rows]
|
||||
|
||||
def get_task_budget(self, task_id: str) -> Budget:
|
||||
|
||||
@@ -0,0 +1,46 @@
|
||||
# External agent ingress. TLS termination is assumed on the enclosing server.
|
||||
# Control endpoints (/v1/agents/register, POST /v1/tasks, verify/cancel/events)
|
||||
# are intentionally NOT exposed here; the AIOS control plane reaches them on
|
||||
# loopback with its server-side credential.
|
||||
|
||||
location = /agent-tasks/health {
|
||||
rewrite ^/agent-tasks/(.*)$ /$1 break;
|
||||
proxy_pass http://127.0.0.1:18940;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
}
|
||||
|
||||
location = /agent-tasks/v1/tasks/poll {
|
||||
limit_except POST { deny all; }
|
||||
rewrite ^/agent-tasks/(.*)$ /$1 break;
|
||||
proxy_pass http://127.0.0.1:18940;
|
||||
proxy_http_version 1.1;
|
||||
proxy_set_header Authorization $http_authorization;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
client_max_body_size 1m;
|
||||
proxy_read_timeout 35s;
|
||||
}
|
||||
|
||||
location ~ ^/agent-tasks/v1/tasks/[A-Za-z0-9][A-Za-z0-9._:-]{2,179}/(claim|heartbeat|result)$ {
|
||||
limit_except POST { deny all; }
|
||||
rewrite ^/agent-tasks/(.*)$ /$1 break;
|
||||
proxy_pass http://127.0.0.1:18940;
|
||||
proxy_http_version 1.1;
|
||||
proxy_set_header Authorization $http_authorization;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
client_max_body_size 1m;
|
||||
proxy_read_timeout 35s;
|
||||
}
|
||||
|
||||
location ~ ^/agent-tasks/v1/tasks/[A-Za-z0-9][A-Za-z0-9._:-]{2,179}$ {
|
||||
limit_except GET { deny all; }
|
||||
rewrite ^/agent-tasks/(.*)$ /$1 break;
|
||||
proxy_pass http://127.0.0.1:18940;
|
||||
proxy_http_version 1.1;
|
||||
proxy_set_header Authorization $http_authorization;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
proxy_read_timeout 35s;
|
||||
}
|
||||
@@ -0,0 +1,36 @@
|
||||
[Unit]
|
||||
Description=AIOS Online Agent Task API
|
||||
After=network-online.target
|
||||
Wants=network-online.target
|
||||
|
||||
[Service]
|
||||
Type=simple
|
||||
User=aios-agent-task
|
||||
Group=aios-agent-task
|
||||
WorkingDirectory=@RUNTIME_ROOT@
|
||||
Environment=PYTHONUNBUFFERED=1
|
||||
Environment=PYTHONDONTWRITEBYTECODE=1
|
||||
Environment=PYTHONPATH=@RUNTIME_ROOT@/app
|
||||
Environment=AIOS_AGENT_TASK_DB=/var/lib/aios-agent-task/state.sqlite3
|
||||
Environment=AIOS_AGENT_TASK_HOST=127.0.0.1
|
||||
Environment=AIOS_AGENT_TASK_PORT=18940
|
||||
Environment=AIOS_AGENT_TASK_CONTROL_TOKEN_FILE=/run/credentials/aios-agent-task-api.service/control_token
|
||||
LoadCredential=control_token:/etc/aios-gitops/agent-task-control.token
|
||||
ExecStart=/usr/bin/python3 @RUNTIME_ROOT@/app/aios_agent_task_api.py
|
||||
Restart=on-failure
|
||||
RestartSec=2
|
||||
NoNewPrivileges=true
|
||||
PrivateTmp=true
|
||||
ProtectHome=true
|
||||
ProtectSystem=strict
|
||||
ProtectKernelTunables=true
|
||||
ProtectKernelModules=true
|
||||
ProtectControlGroups=true
|
||||
LockPersonality=true
|
||||
RestrictSUIDSGID=true
|
||||
RestrictAddressFamilies=AF_UNIX AF_INET AF_INET6
|
||||
ReadOnlyPaths=@RUNTIME_ROOT@
|
||||
ReadWritePaths=/var/lib/aios-agent-task
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,7 @@
|
||||
# Existing model API drop-in. Install only with the governed runtime package.
|
||||
[Service]
|
||||
SupplementaryGroups=gitops aios-executor-exchange aios-autonomous-agent-worker
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_SOCKET=/run/aios-owner-authorization-broker/broker.sock
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_ACCOUNT=chatgpt-governed-implementer-singlehost
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_CALLER=chatgpt-governed-implementer-singlehost
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_SCOPE=release:candidate
|
||||
@@ -28,6 +28,9 @@ Environment=AIOS_GITOPS_MODEL_API_SERVICE=aios-gitops-model-api-v2r2c1-staging.s
|
||||
Environment=AIOS_SILENT_RELEASE_EXECUTION_ENABLED=false
|
||||
Environment=AIOS_SILENT_RELEASE_STATE_ROOT=/var/lib/aios-gitops/model-api-staging/silent-release
|
||||
Environment=AIOS_BRIDGE_V2_URL=http://127.0.0.1:18892
|
||||
Environment=AIOS_AGENT_TASK_API_URL=http://127.0.0.1:18940
|
||||
Environment=AIOS_AGENT_TASK_CONTROL_TOKEN_FILE=/run/credentials/aios-gitops-model-api-v2r2c1-staging.service/agent_task_control_token
|
||||
LoadCredential=agent_task_control_token:/etc/aios-gitops/agent-task-control.token
|
||||
Environment=AIOS_GITOPS_HOST=127.0.0.1
|
||||
Environment=AIOS_GITOPS_PORT=18893
|
||||
Environment=PYTHONPATH=@RUNTIME_ROOT@/app:@RUNTIME_ROOT@/runtime-agent-os
|
||||
|
||||
@@ -0,0 +1,49 @@
|
||||
[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_ALLOWED_USERS=aios-autonomous-agent-worker,aiman
|
||||
Environment=AIOS_OWNER_AUTH_BROKER_SOCKET=%t/aios-owner-authorization-broker/broker.sock
|
||||
ExecStart=/usr/bin/python3 -B @RUNTIME_ROOT@/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
|
||||
ReadOnlyPaths=@RUNTIME_ROOT@
|
||||
InaccessiblePaths=/data/aios-gitops /home
|
||||
UMask=0007
|
||||
|
||||
[Install]
|
||||
WantedBy=multi-user.target
|
||||
@@ -0,0 +1,88 @@
|
||||
# Caller admission projection and rehydration
|
||||
|
||||
This runbook covers the exact durable Owner policy and the exact API caller:
|
||||
|
||||
- caller: \`chatgpt-governed-implementer-singlehost\`
|
||||
- scope: \`release:candidate\`
|
||||
- policy source: \`/etc/aios-gitops/owner-durable-authorization.json\`
|
||||
- revocation source: \`/etc/aios-gitops/owner-durable-authorization-revocations.json\`
|
||||
|
||||
## Admission source
|
||||
|
||||
The model API first authenticates the existing caller token from the static
|
||||
registry. For this exact caller and exact release scope, static authentication
|
||||
is not sufficient. The model API calls the existing Unix-socket Owner broker.
|
||||
The broker reloads and validates the durable policy on every projection request,
|
||||
including account, capability, optional expiry, and revocation.
|
||||
|
||||
Other callers retain their existing static-scope behavior. The target caller
|
||||
does not receive a permanent \`release:candidate\` entry in the static registry,
|
||||
and it is never treated as a release-worker account.
|
||||
|
||||
## Projection lifecycle and ACK semantics
|
||||
|
||||
The broker records only non-secret projection metadata in its existing durable
|
||||
state database. The state sequence is:
|
||||
|
||||
\`PENDING\` -> \`ACKED\`
|
||||
|
||||
The ACK is produced by the broker's server-side projection consumer after the
|
||||
exact policy and request binding have passed validation. The public response
|
||||
contains projection state, effective scope, policy digest, expiry and
|
||||
rehydration state. The short-lived credential material is returned only in a
|
||||
private Unix-socket response header to the model API process. It is not written
|
||||
to the response body, logs, diagnostics, Git, task metadata or terminal output.
|
||||
|
||||
## Credential rehydration, expiry and revocation
|
||||
|
||||
The credential projection is generated in memory for a short TTL and is
|
||||
discarded after the model API admission check. The durable row stores no raw
|
||||
credential and no bearer token. After a broker or model API restart, the next
|
||||
admission request creates a new short-lived credential from the durable
|
||||
projection metadata; this is reported as \`RECOVERED\`.
|
||||
|
||||
Every request reloads the policy. Missing, malformed, expired or revoked policy
|
||||
fails closed. A caller mismatch, scope mismatch, invalid projection state,
|
||||
broker outage, missing credential header or unexpected secret field also fails
|
||||
closed. No static scope fallback is allowed for the target caller.
|
||||
|
||||
## Diagnosis
|
||||
|
||||
Use the read-only API with the caller's existing credential:
|
||||
|
||||
\`GET /api/caller-capabilities\`
|
||||
|
||||
The response is safe to display and contains:
|
||||
|
||||
- \`caller_identity\`
|
||||
- \`admitted\`
|
||||
- \`effective_scopes\`
|
||||
- \`projection_state\`
|
||||
- \`credential_state\`
|
||||
- \`rehydration_state\`
|
||||
- \`secret_exposed=false\`
|
||||
|
||||
It never contains a bearer, token, credential value or credential hash.
|
||||
|
||||
For a release candidate request, the model API also binds the exact
|
||||
\`source_commit\` and \`version\` to the durable policy before invoking the
|
||||
trusted release service. The caller cannot supply release evidence fields.
|
||||
|
||||
## Failure states and troubleshooting
|
||||
|
||||
- \`PROJECTION_BROKER_UNAVAILABLE\`: verify the broker unit, socket path and
|
||||
the model API supplementary groups; do not add a static scope.
|
||||
- \`OWNER_GRANT_REVOKED\` or \`POLICY_EXPIRED\`: restore a valid Owner policy
|
||||
through governed policy management; do not edit broker state.
|
||||
- \`CALLER_MISMATCH\` or \`SCOPE_MISMATCH\`: verify the authenticated account and
|
||||
requested exact scope.
|
||||
- \`PROJECTION_STATE_INVALID\`: inspect the durable projection through the
|
||||
broker's read-only diagnostics and fail closed until it is reconciled.
|
||||
- \`CREDENTIAL_REHYDRATION_MISSING\`: restart only the related control-plane
|
||||
service after capturing pre-state and then repeat the read-only diagnosis.
|
||||
- \`trusted candidate builder is not configured\`: Gate A admission is working,
|
||||
but the separate existing TrustedCandidateBuilder runtime must be restored
|
||||
before Gate B can create a candidate.
|
||||
|
||||
No Owner Approval, Deployment, Host Seed, Host Activation, Projector,
|
||||
continuation task, Tag or Baseline action is part of this path.
|
||||
@@ -0,0 +1,295 @@
|
||||
openapi: 3.1.0
|
||||
info:
|
||||
title: AIOS Online Agent Task API
|
||||
version: 1.0.0
|
||||
description: >
|
||||
Durable online task distribution for OpenClaw, Hermes, and other external
|
||||
agents. Agent credentials and lease tokens are bearer credentials and must
|
||||
never be committed to Git.
|
||||
servers:
|
||||
- url: /agent-tasks
|
||||
paths:
|
||||
/health:
|
||||
get:
|
||||
operationId: agentTaskHealth
|
||||
responses:
|
||||
"200":
|
||||
description: Service health
|
||||
/v1/agents/register:
|
||||
post:
|
||||
operationId: registerOnlineAgent
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Control-plane only. Returns the agent token once.
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/RegisterAgentRequest"
|
||||
responses:
|
||||
"201":
|
||||
description: Registered
|
||||
/v1/tasks:
|
||||
post:
|
||||
operationId: createOnlineAgentTask
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Control-plane only.
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/CreateTaskRequest"
|
||||
responses:
|
||||
"201":
|
||||
description: Task created
|
||||
/v1/tasks/poll:
|
||||
post:
|
||||
operationId: pollOnlineAgentTasks
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Agent operation. Returns capability-compatible queued tasks.
|
||||
requestBody:
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
properties:
|
||||
limit:
|
||||
type: integer
|
||||
minimum: 1
|
||||
maximum: 100
|
||||
default: 20
|
||||
responses:
|
||||
"200":
|
||||
description: Available tasks
|
||||
/v1/tasks/{task_id}:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
get:
|
||||
operationId: getOnlineAgentTaskDirect
|
||||
security:
|
||||
- bearerAuth: []
|
||||
responses:
|
||||
"200":
|
||||
description: Task status
|
||||
/v1/tasks/{task_id}/claim:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
post:
|
||||
operationId: claimOnlineAgentTask
|
||||
security:
|
||||
- bearerAuth: []
|
||||
requestBody:
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
properties:
|
||||
lease_seconds:
|
||||
type: integer
|
||||
minimum: 30
|
||||
maximum: 3600
|
||||
default: 300
|
||||
responses:
|
||||
"200":
|
||||
description: >
|
||||
Claim succeeded. The lease_token is returned once and must be kept
|
||||
in runtime memory/secret storage, never Git.
|
||||
"409":
|
||||
description: Already claimed or otherwise not claimable
|
||||
/v1/tasks/{task_id}/heartbeat:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
post:
|
||||
operationId: heartbeatOnlineAgentTask
|
||||
security:
|
||||
- bearerAuth: []
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/HeartbeatRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Lease renewed
|
||||
/v1/tasks/{task_id}/result:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
post:
|
||||
operationId: submitOnlineAgentResult
|
||||
security:
|
||||
- bearerAuth: []
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/ResultRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Result accepted into VERIFYING
|
||||
/v1/tasks/{task_id}/verify:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
post:
|
||||
operationId: verifyOnlineAgentTaskDirect
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Control-plane only.
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/VerifyRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Verification decision applied
|
||||
/v1/tasks/{task_id}/cancel:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
post:
|
||||
operationId: cancelOnlineAgentTaskDirect
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Control-plane only.
|
||||
requestBody:
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
properties:
|
||||
reason:
|
||||
type: string
|
||||
responses:
|
||||
"200":
|
||||
description: Cancelled or already terminal
|
||||
/v1/tasks/{task_id}/events:
|
||||
parameters:
|
||||
- $ref: "#/components/parameters/TaskId"
|
||||
get:
|
||||
operationId: getOnlineAgentTaskEventsDirect
|
||||
security:
|
||||
- bearerAuth: []
|
||||
description: Control-plane only.
|
||||
responses:
|
||||
"200":
|
||||
description: Audit event list
|
||||
components:
|
||||
securitySchemes:
|
||||
bearerAuth:
|
||||
type: http
|
||||
scheme: bearer
|
||||
parameters:
|
||||
TaskId:
|
||||
name: task_id
|
||||
in: path
|
||||
required: true
|
||||
schema:
|
||||
type: string
|
||||
pattern: "^[A-Za-z0-9][A-Za-z0-9._:-]{2,179}$"
|
||||
schemas:
|
||||
RegisterAgentRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [agent_id, agent_type, agent_version, capabilities]
|
||||
properties:
|
||||
agent_id:
|
||||
type: string
|
||||
agent_type:
|
||||
type: string
|
||||
examples: [openclaw]
|
||||
agent_version:
|
||||
type: string
|
||||
capabilities:
|
||||
type: array
|
||||
items:
|
||||
type: string
|
||||
metadata:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
CreateTaskRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required:
|
||||
- goal
|
||||
- instructions
|
||||
- success_criteria
|
||||
- idempotency_key
|
||||
properties:
|
||||
task_id:
|
||||
type: string
|
||||
goal:
|
||||
type: string
|
||||
instructions:
|
||||
type: string
|
||||
success_criteria:
|
||||
type: array
|
||||
items: {}
|
||||
required_capabilities:
|
||||
type: array
|
||||
items:
|
||||
type: string
|
||||
constraints:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
metadata:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
deadline:
|
||||
type: string
|
||||
format: date-time
|
||||
priority:
|
||||
type: integer
|
||||
minimum: -100
|
||||
maximum: 100
|
||||
default: 0
|
||||
max_attempts:
|
||||
type: integer
|
||||
minimum: 1
|
||||
maximum: 20
|
||||
default: 3
|
||||
idempotency_key:
|
||||
type: string
|
||||
HeartbeatRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [lease_token]
|
||||
properties:
|
||||
lease_token:
|
||||
type: string
|
||||
lease_seconds:
|
||||
type: integer
|
||||
minimum: 30
|
||||
maximum: 3600
|
||||
default: 300
|
||||
progress:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
ResultRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [lease_token, idempotency_key, result]
|
||||
properties:
|
||||
lease_token:
|
||||
type: string
|
||||
idempotency_key:
|
||||
type: string
|
||||
result:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
VerifyRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [decision]
|
||||
properties:
|
||||
decision:
|
||||
type: string
|
||||
enum: [SUCCEEDED, NEEDS_REWORK, FAILED]
|
||||
verification:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,326 @@
|
||||
openapi: 3.1.0
|
||||
info:
|
||||
title: AIOS ChatGPT Governed Write and Agent Task Extension
|
||||
version: 1.0.0
|
||||
description: >
|
||||
Merge these paths and schemas into docs/gitops/openapi-chatgpt-actions.yaml.
|
||||
This extension deliberately exposes governed Bridge v2 operations rather
|
||||
than reopening legacy AIOS_GITOPS_WRITES_ENABLED.
|
||||
paths:
|
||||
/api/governed-write-capabilities:
|
||||
get:
|
||||
operationId: getGovernedWriteCapabilities
|
||||
responses:
|
||||
"200":
|
||||
description: Governed and legacy write capability status
|
||||
/api/control-plane-snapshot:
|
||||
get:
|
||||
operationId: getControlPlaneSnapshot
|
||||
responses:
|
||||
"200":
|
||||
description: Non-sensitive control-plane snapshot
|
||||
/createGovernedTask:
|
||||
post:
|
||||
operationId: createGovernedTask
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/CreateGovernedTaskRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Governed task binding
|
||||
/acquireWriterLease:
|
||||
post:
|
||||
operationId: acquireWriterLease
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/WriterLeaseRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Writer lease binding
|
||||
/commitFileGoverned:
|
||||
post:
|
||||
operationId: commitFileGoverned
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/CommitFileGovernedRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Durable GitOps job receipt
|
||||
/commitFilesGoverned:
|
||||
post:
|
||||
operationId: commitFilesGoverned
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/CommitFilesGovernedRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Durable GitOps job receipt
|
||||
/getGitOpsJob:
|
||||
post:
|
||||
operationId: getGitOpsJob
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [job_id]
|
||||
properties:
|
||||
job_id: {type: string}
|
||||
responses:
|
||||
"200":
|
||||
description: Job receipt
|
||||
/resumeGitOpsJob:
|
||||
post:
|
||||
operationId: resumeGitOpsJob
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [job_id, expected_revision]
|
||||
properties:
|
||||
job_id: {type: string}
|
||||
expected_revision: {type: integer}
|
||||
responses:
|
||||
"200":
|
||||
description: Resumed job receipt
|
||||
/reconcileGitOpsJob:
|
||||
post:
|
||||
operationId: reconcileGitOpsJob
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [job_id, expected_revision]
|
||||
properties:
|
||||
job_id: {type: string}
|
||||
expected_revision: {type: integer}
|
||||
responses:
|
||||
"200":
|
||||
description: Reconciliation receipt
|
||||
/createOnlineAgentTask:
|
||||
post:
|
||||
operationId: createOnlineAgentTask
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: "#/components/schemas/CreateOnlineAgentTaskRequest"
|
||||
responses:
|
||||
"200":
|
||||
description: Online task created
|
||||
"201":
|
||||
description: Online task created
|
||||
/registerOnlineAgent:
|
||||
post:
|
||||
operationId: registerOnlineAgent
|
||||
x-openai-isConsequential: true
|
||||
description: >
|
||||
Control operation. The returned agent token is secret material and must
|
||||
be delivered through runtime secret storage, never Git.
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [agent_id, agent_type, agent_version, capabilities]
|
||||
properties:
|
||||
agent_id: {type: string}
|
||||
agent_type: {type: string}
|
||||
agent_version: {type: string}
|
||||
capabilities:
|
||||
type: array
|
||||
items: {type: string}
|
||||
metadata:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
responses:
|
||||
"201":
|
||||
description: Agent registered
|
||||
/api/online-agent-task:
|
||||
get:
|
||||
operationId: getOnlineAgentTask
|
||||
parameters:
|
||||
- name: task_id
|
||||
in: query
|
||||
required: true
|
||||
schema: {type: string}
|
||||
responses:
|
||||
"200":
|
||||
description: Task status
|
||||
/api/online-agent-task-events:
|
||||
get:
|
||||
operationId: getOnlineAgentTaskEvents
|
||||
parameters:
|
||||
- name: task_id
|
||||
in: query
|
||||
required: true
|
||||
schema: {type: string}
|
||||
responses:
|
||||
"200":
|
||||
description: Task events
|
||||
/verifyOnlineAgentTask:
|
||||
post:
|
||||
operationId: verifyOnlineAgentTask
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [task_id, decision]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
decision:
|
||||
type: string
|
||||
enum: [SUCCEEDED, NEEDS_REWORK, FAILED]
|
||||
verification:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
responses:
|
||||
"200":
|
||||
description: Verification applied
|
||||
/cancelOnlineAgentTask:
|
||||
post:
|
||||
operationId: cancelOnlineAgentTask
|
||||
x-openai-isConsequential: true
|
||||
requestBody:
|
||||
required: true
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: object
|
||||
required: [task_id]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
reason: {type: string}
|
||||
responses:
|
||||
"200":
|
||||
description: Task cancelled
|
||||
components:
|
||||
schemas:
|
||||
Oid40:
|
||||
type: string
|
||||
pattern: "^[0-9a-f]{40}$"
|
||||
CreateGovernedTaskRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required:
|
||||
[task_id, repository, base_oid, target_branch, risk, description, idempotency_key]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
repository: {type: string, const: aios/aios}
|
||||
base_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
target_branch:
|
||||
type: string
|
||||
pattern: "^agent/"
|
||||
risk:
|
||||
type: string
|
||||
enum: [low, medium, high, engineering]
|
||||
description: {type: string}
|
||||
idempotency_key: {type: string}
|
||||
expected_branch_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
generation: {type: integer, minimum: 1}
|
||||
WriterLeaseRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required:
|
||||
[schema, task_id, repository, branch_mode, base_oid, expected_branch_oid,
|
||||
target_branch, generation, lease_seconds]
|
||||
properties:
|
||||
schema: {type: string}
|
||||
task_id: {type: string}
|
||||
repository: {type: string, const: aios/aios}
|
||||
branch_mode:
|
||||
type: string
|
||||
enum: [CREATE_NEW, CONTINUE_EXISTING]
|
||||
base_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
expected_branch_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
target_branch: {type: string, pattern: "^agent/"}
|
||||
generation: {type: integer, minimum: 1}
|
||||
lease_seconds: {type: integer, minimum: 1, maximum: 86400}
|
||||
CommitFileGovernedRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required:
|
||||
[task_id, repository, expected_branch_oid, target_branch, generation,
|
||||
idempotency_key, path, content_base64, message]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
repository: {type: string, const: aios/aios}
|
||||
expected_branch_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
target_branch: {type: string, pattern: "^agent/"}
|
||||
generation: {type: integer, minimum: 1}
|
||||
idempotency_key: {type: string}
|
||||
path: {type: string}
|
||||
content_base64: {type: string}
|
||||
message: {type: string}
|
||||
CommitFilesGovernedRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required:
|
||||
[task_id, repository, expected_branch_oid, target_branch, generation,
|
||||
idempotency_key, files, message]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
repository: {type: string, const: aios/aios}
|
||||
expected_branch_oid: {$ref: "#/components/schemas/Oid40"}
|
||||
target_branch: {type: string, pattern: "^agent/"}
|
||||
generation: {type: integer, minimum: 1}
|
||||
idempotency_key: {type: string}
|
||||
files:
|
||||
type: object
|
||||
additionalProperties:
|
||||
type: string
|
||||
description: Base64 encoded file bytes
|
||||
message: {type: string}
|
||||
CreateOnlineAgentTaskRequest:
|
||||
type: object
|
||||
additionalProperties: false
|
||||
required: [goal, instructions, success_criteria, idempotency_key]
|
||||
properties:
|
||||
task_id: {type: string}
|
||||
goal: {type: string}
|
||||
instructions: {type: string}
|
||||
success_criteria:
|
||||
type: array
|
||||
items: {}
|
||||
required_capabilities:
|
||||
type: array
|
||||
items: {type: string}
|
||||
constraints:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
metadata:
|
||||
type: object
|
||||
additionalProperties: true
|
||||
deadline: {type: string, format: date-time}
|
||||
priority: {type: integer, minimum: -100, maximum: 100}
|
||||
max_attempts: {type: integer, minimum: 1, maximum: 20}
|
||||
idempotency_key: {type: string}
|
||||
@@ -1,54 +0,0 @@
|
||||
# R4.2 P0-A TaskPlane Scope Isolation Design
|
||||
|
||||
Status: approved candidate design. This document is not evidence of a protected-main merge or production deployment.
|
||||
|
||||
## Boundary
|
||||
|
||||
- Task: `r42-p0a-taskplane-scope-isolation-20260816`
|
||||
- Governed Job: `job-3fd743b51971c6007923699c1233f84c`
|
||||
- Repository: `aios/aios`
|
||||
- Base: `0eb6e1a864edfdaeef37b438991707a110e21558`
|
||||
- Target branch: `agent/r42-p0a-taskplane-scope-isolation-20260816`
|
||||
- Generation: `1`
|
||||
|
||||
The implementation reuses this exact boundary. It does not create a replacement Task, Job, branch, generation, or idempotency key.
|
||||
|
||||
## Model and invariants
|
||||
|
||||
`TaskPlan` gains first-class `scope_type`, `tenant_id`, `user_id`, `project_id`, `conversation_id`, and `run_id` fields. The fields participate in accepted-key validation, parsing, serialization, persistence, and authorization.
|
||||
|
||||
Historical rows without scope data normalize to `scope_type=legacy`; no tenant is inferred. Tenant-scoped tasks require a non-empty `tenant_id`. System/global resources use an explicit `scope_type=system` invariant and do not receive a fabricated tenant. A task may add user, project, conversation, and run ownership within its tenant.
|
||||
|
||||
## Trusted ActorContext
|
||||
|
||||
TaskPlane defines a small immutable `TaskActorContext` used only for authorization. The HTTP authenticator returns an identity. `PolicyEngine` resolves that identity through server-owned policy configuration into the effective context and capabilities. Request fields such as `tenant_id`, `user_id`, `group_id`, and `role` are declarations only and never expand authority.
|
||||
|
||||
Public create, list, and get paths always use the resolved context. Internal compatibility calls may omit a context and retain explicit legacy behavior, but the HTTP adapter never uses that path.
|
||||
|
||||
## Persistence
|
||||
|
||||
SQLite remains the current runtime. `_create_schema()` performs an idempotent forward migration that adds the six scope columns and indexes, then backfills only missing `scope_type` values to `legacy`.
|
||||
|
||||
New writes persist scope in both the canonical `task_json` representation and first-class columns. Reads accept historical JSON without scope, using legacy columns, but reject a new-row mismatch between JSON and columns. SQL list/get predicates are parameterized and applied before row hydration.
|
||||
|
||||
PostgreSQL remains a separate migration target and is not claimed by this change.
|
||||
|
||||
## Authorization flow
|
||||
|
||||
1. HTTP bearer authentication yields an identity.
|
||||
2. `PolicyEngine` verifies capability and resolves the identity to `TaskActorContext`.
|
||||
3. Create binds the requested scope to the trusted context and rejects mismatches.
|
||||
4. List/get pass the trusted context to `TaskPlaneStore`.
|
||||
5. Store applies tenant and ownership predicates in SQL; privileged capabilities may explicitly cross scope.
|
||||
|
||||
Tenant actors cannot read legacy tasks, another tenant, or mismatched owned resources. Legacy actors can only read legacy tasks. System actors require explicit privileged capability for cross-scope access.
|
||||
|
||||
## Failure behavior
|
||||
|
||||
Missing tenant scope, forged scope, unknown scope types, JSON/column divergence, and unauthorized list/get operations fail closed with canonical TaskPlane errors. Public get does not reveal whether an out-of-scope task exists.
|
||||
|
||||
## Verification
|
||||
|
||||
Tests exercise real model round trips, SQLite migration/restart, Service/Policy/Store wiring, tenant A versus tenant B list/get denial, ownership filtering, forged create payload rejection, legacy isolation, privileged access, and SQL parameterization. Existing TaskPlane API, RBAC, recovery/security, and security-contract suites remain mandatory.
|
||||
|
||||
The candidate is complete only after real `TASK_CLAIMED`, `TASK_HEARTBEAT`, exact WriterLease, governed commit, OID convergence, result submission, and `TASK_VERIFIED`. Protected main and production remain unchanged.
|
||||
@@ -1,9 +0,0 @@
|
||||
# R4.2 P0-A TaskPlane Scope Isolation
|
||||
|
||||
Status: candidate implementation. This is not proof of protected-main promotion or production deployment.
|
||||
|
||||
Task scope is now represented by `scope_type`, `tenant_id`, `user_id`, `project_id`, `conversation_id`, and `run_id` in the model and SQLite persistence. Historical rows become `legacy`; their tenant is never guessed. Tenant tasks require `tenant_id`. System resources use `scope_type=system` without a fabricated tenant.
|
||||
|
||||
The HTTP adapter uses the authenticated identity. `PolicyEngine` resolves that identity through server-owned `identity_actor_contexts`; request payload fields do not grant scope. Public create binds the trusted scope, and public list/get apply parameterized scope predicates in `TaskPlaneStore`.
|
||||
|
||||
SQLite is the current runtime database. PostgreSQL remains a separate migration and runtime-acceptance gate.
|
||||
@@ -0,0 +1,20 @@
|
||||
{
|
||||
"execution_id": "exec_d75446452723",
|
||||
"user_id": "user1",
|
||||
"visible_scope": "group1",
|
||||
"entry_agent": "openclaw",
|
||||
"path": [
|
||||
"openclaw",
|
||||
"hermes",
|
||||
"codex"
|
||||
],
|
||||
"used_hermes": true,
|
||||
"closed_by": "codex",
|
||||
"duration_ms": 0,
|
||||
"input_tokens": 100,
|
||||
"output_tokens": 200,
|
||||
"total_tokens": 300,
|
||||
"status": "completed",
|
||||
"started_at": "2026-08-09T17:13:34.996470+00:00",
|
||||
"ended_at": "2026-08-09T17:13:34.996489+00:00"
|
||||
}
|
||||
@@ -0,0 +1,2 @@
|
||||
{"task_id": "task-command_watcher-CMD-test", "title": "CMD-test", "assigned_by": "command-watcher", "executor": "openclaw", "source": "command_watcher", "status": "pending", "priority": "normal", "created_at": "2026-08-09T17:13:47.285193+00:00", "updated_at": "2026-08-09T17:13:47.285198+00:00", "progress": {"percent": 5, "current_step": "pending", "completed_steps": [], "remaining_steps": [], "blocked_reason": null}, "notification": {"required": true, "sent": false, "channel": null, "sent_at": null}, "evidence": ["ops/openclaw/CMD-test.md"], "metadata": {"bucket": "active"}}
|
||||
{"task_id": "task-agent_execution-e1", "title": "Agent execution e1", "assigned_by": "hermes", "executor": "hermes", "source": "agent_execution", "status": "completed", "priority": "normal", "created_at": "2026-08-09T17:13:47.285822+00:00", "updated_at": "2026-08-09T17:13:47.285826+00:00", "progress": {"percent": 100, "current_step": "completed", "completed_steps": [], "remaining_steps": [], "blocked_reason": null}, "notification": {"required": true, "sent": false, "channel": null, "sent_at": null}, "evidence": ["e1"], "metadata": {"execution_id": "e1", "entry_agent": "hermes", "status": "completed"}}
|
||||
@@ -0,0 +1,112 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Small CLI used by OpenClaw/Hermes operators to validate AIOS online tasks."""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import json
|
||||
from pathlib import Path
|
||||
import sys
|
||||
from urllib.error import HTTPError, URLError
|
||||
from urllib.request import Request, urlopen
|
||||
|
||||
|
||||
def read_token(value: str | None, path: str | None) -> str:
|
||||
if value:
|
||||
return value.strip()
|
||||
if path:
|
||||
return Path(path).read_text(encoding="utf-8").strip()
|
||||
raise SystemExit("token or token-file is required")
|
||||
|
||||
|
||||
def request(base: str, method: str, path: str, token: str, body=None):
|
||||
payload = None if body is None else json.dumps(body, ensure_ascii=False).encode("utf-8")
|
||||
req = Request(
|
||||
base.rstrip("/") + path,
|
||||
data=payload,
|
||||
method=method,
|
||||
headers={
|
||||
"Authorization": "Bearer " + token,
|
||||
"Accept": "application/json",
|
||||
**({"Content-Type": "application/json"} if payload is not None else {}),
|
||||
},
|
||||
)
|
||||
try:
|
||||
with urlopen(req, timeout=30) as response:
|
||||
return response.status, json.loads(response.read().decode("utf-8"))
|
||||
except HTTPError as exc:
|
||||
raw = exc.read()
|
||||
try:
|
||||
value = json.loads(raw.decode("utf-8"))
|
||||
except Exception:
|
||||
value = {"error": "HTTP_" + str(exc.code)}
|
||||
return exc.code, value
|
||||
except (URLError, TimeoutError) as exc:
|
||||
raise SystemExit("network_error:" + type(exc).__name__) from exc
|
||||
|
||||
|
||||
def main() -> int:
|
||||
ap = argparse.ArgumentParser()
|
||||
ap.add_argument("--base", required=True)
|
||||
ap.add_argument("--token")
|
||||
ap.add_argument("--token-file")
|
||||
sub = ap.add_subparsers(dest="command", required=True)
|
||||
|
||||
p = sub.add_parser("poll")
|
||||
p.add_argument("--limit", type=int, default=20)
|
||||
|
||||
p = sub.add_parser("claim")
|
||||
p.add_argument("task_id")
|
||||
p.add_argument("--lease-seconds", type=int, default=300)
|
||||
|
||||
p = sub.add_parser("heartbeat")
|
||||
p.add_argument("task_id")
|
||||
p.add_argument("--lease-token", required=True)
|
||||
p.add_argument("--lease-seconds", type=int, default=300)
|
||||
p.add_argument("--progress-json", default="{}")
|
||||
|
||||
p = sub.add_parser("result")
|
||||
p.add_argument("task_id")
|
||||
p.add_argument("--lease-token", required=True)
|
||||
p.add_argument("--idempotency-key", required=True)
|
||||
p.add_argument("--result-json", required=True)
|
||||
|
||||
p = sub.add_parser("status")
|
||||
p.add_argument("task_id")
|
||||
|
||||
args = ap.parse_args()
|
||||
token = read_token(args.token, args.token_file)
|
||||
|
||||
if args.command == "poll":
|
||||
status, value = request(args.base, "POST", "/v1/tasks/poll", token, {"limit": args.limit})
|
||||
elif args.command == "claim":
|
||||
status, value = request(
|
||||
args.base, "POST", f"/v1/tasks/{args.task_id}/claim", token,
|
||||
{"lease_seconds": args.lease_seconds},
|
||||
)
|
||||
elif args.command == "heartbeat":
|
||||
status, value = request(
|
||||
args.base, "POST", f"/v1/tasks/{args.task_id}/heartbeat", token,
|
||||
{
|
||||
"lease_token": args.lease_token,
|
||||
"lease_seconds": args.lease_seconds,
|
||||
"progress": json.loads(args.progress_json),
|
||||
},
|
||||
)
|
||||
elif args.command == "result":
|
||||
status, value = request(
|
||||
args.base, "POST", f"/v1/tasks/{args.task_id}/result", token,
|
||||
{
|
||||
"lease_token": args.lease_token,
|
||||
"idempotency_key": args.idempotency_key,
|
||||
"result": json.loads(args.result_json),
|
||||
},
|
||||
)
|
||||
else:
|
||||
status, value = request(args.base, "GET", f"/v1/tasks/{args.task_id}", token)
|
||||
|
||||
print(json.dumps({"http_status": status, "response": value}, ensure_ascii=False, indent=2))
|
||||
return 0 if 200 <= status < 300 else 1
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -0,0 +1,193 @@
|
||||
#!/usr/bin/env python3
|
||||
"""End-to-end online task feasibility smoke test.
|
||||
|
||||
Runs through the public agent path plus control operations:
|
||||
create -> poll -> claim -> heartbeat -> result -> verify -> readback.
|
||||
It never performs Git, tag, Baseline, or production mutations.
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
from datetime import datetime, timezone
|
||||
import json
|
||||
from pathlib import Path
|
||||
import secrets
|
||||
from urllib.error import HTTPError, URLError
|
||||
from urllib.request import Request, urlopen
|
||||
|
||||
|
||||
def token(path: Path) -> str:
|
||||
value = path.read_text(encoding="utf-8").strip()
|
||||
if len(value) < 32:
|
||||
raise SystemExit("credential_invalid:" + str(path))
|
||||
return value
|
||||
|
||||
|
||||
def call(base: str, method: str, route: str, credential: str, body=None):
|
||||
payload = None if body is None else json.dumps(body, ensure_ascii=False).encode("utf-8")
|
||||
headers = {
|
||||
"Authorization": "Bearer " + credential,
|
||||
"Accept": "application/json",
|
||||
}
|
||||
if payload is not None:
|
||||
headers["Content-Type"] = "application/json"
|
||||
request = Request(base.rstrip("/") + route, data=payload, headers=headers, method=method)
|
||||
try:
|
||||
with urlopen(request, timeout=30) as response:
|
||||
return response.status, json.loads(response.read().decode("utf-8"))
|
||||
except HTTPError as exc:
|
||||
try:
|
||||
value = json.loads(exc.read().decode("utf-8"))
|
||||
except Exception:
|
||||
value = {"error": "HTTP_" + str(exc.code)}
|
||||
return exc.code, value
|
||||
except (URLError, TimeoutError) as exc:
|
||||
raise SystemExit("network_error:" + type(exc).__name__) from exc
|
||||
|
||||
|
||||
def require(step: str, status: int, value, expected_status: str | None = None):
|
||||
if not 200 <= status < 300:
|
||||
raise SystemExit(f"{step}_failed:http={status}:body={json.dumps(value, ensure_ascii=False)}")
|
||||
if expected_status and value.get("status") != expected_status:
|
||||
raise SystemExit(
|
||||
f"{step}_state_failed:expected={expected_status}:"
|
||||
f"observed={value.get('status')}"
|
||||
)
|
||||
print(json.dumps({"step": step, "http": status, "response": value}, ensure_ascii=False))
|
||||
|
||||
|
||||
def main() -> int:
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("--control-base", required=True,
|
||||
help="Loopback/model-control base that exposes /v1 control routes")
|
||||
parser.add_argument("--agent-base", required=True,
|
||||
help="Public agent base, e.g. https://host/agent-tasks")
|
||||
parser.add_argument("--control-token-file", required=True, type=Path)
|
||||
parser.add_argument("--agent-token-file", required=True, type=Path)
|
||||
args = parser.parse_args()
|
||||
|
||||
control = token(args.control_token_file)
|
||||
agent = token(args.agent_token_file)
|
||||
task_id = "TASK-ONLINE-SMOKE-" + datetime.now(timezone.utc).strftime("%Y%m%dT%H%M%SZ") + "-" + secrets.token_hex(3)
|
||||
create_key = "smoke-create:" + task_id
|
||||
result_key = "smoke-result:" + task_id
|
||||
|
||||
s, v = call(
|
||||
args.control_base, "POST", "/v1/tasks", control,
|
||||
{
|
||||
"task_id": task_id,
|
||||
"goal": "Validate AIOS online task distribution end to end",
|
||||
"instructions": (
|
||||
"Do not mutate Git or production. Return a structured result "
|
||||
"containing marker ONLINE_TASK_FLOW_OK."
|
||||
),
|
||||
"success_criteria": [
|
||||
{"field": "marker", "equals": "ONLINE_TASK_FLOW_OK"},
|
||||
{"field": "agent_observed_task_id", "equals": task_id},
|
||||
],
|
||||
"required_capabilities": ["online-task-smoke"],
|
||||
"constraints": {
|
||||
"git_mutation": False,
|
||||
"production_mutation": False,
|
||||
"network_scope": "AIOS task endpoint only",
|
||||
},
|
||||
"priority": 100,
|
||||
"max_attempts": 3,
|
||||
"idempotency_key": create_key,
|
||||
},
|
||||
)
|
||||
require("create", s, v, "QUEUED")
|
||||
|
||||
s, v = call(args.agent_base, "POST", "/v1/tasks/poll", agent, {"limit": 100})
|
||||
require("poll", s, v)
|
||||
if task_id not in [item.get("task_id") for item in v.get("tasks", [])]:
|
||||
raise SystemExit("poll_failed:created_task_not_visible")
|
||||
|
||||
s, v = call(
|
||||
args.agent_base, "POST", f"/v1/tasks/{task_id}/claim", agent,
|
||||
{"lease_seconds": 120},
|
||||
)
|
||||
require("claim", s, v, "CLAIMED")
|
||||
lease = v.get("lease_token")
|
||||
if not lease:
|
||||
raise SystemExit("claim_failed:lease_token_missing")
|
||||
|
||||
s, v = call(
|
||||
args.agent_base, "POST", f"/v1/tasks/{task_id}/heartbeat", agent,
|
||||
{
|
||||
"lease_token": lease,
|
||||
"lease_seconds": 120,
|
||||
"progress": {"stage": "executing", "percent": 50},
|
||||
},
|
||||
)
|
||||
require("heartbeat", s, v, "RUNNING")
|
||||
|
||||
s, v = call(
|
||||
args.agent_base, "POST", f"/v1/tasks/{task_id}/result", agent,
|
||||
{
|
||||
"lease_token": lease,
|
||||
"idempotency_key": result_key,
|
||||
"result": {
|
||||
"status": "completed",
|
||||
"marker": "ONLINE_TASK_FLOW_OK",
|
||||
"agent_observed_task_id": task_id,
|
||||
"summary": "Agent received, claimed, heartbeated and returned the online task.",
|
||||
"artifacts": [],
|
||||
},
|
||||
},
|
||||
)
|
||||
require("result", s, v, "VERIFYING")
|
||||
|
||||
# Re-submit the identical result to prove idempotency.
|
||||
s2, v2 = call(
|
||||
args.agent_base, "POST", f"/v1/tasks/{task_id}/result", agent,
|
||||
{
|
||||
"lease_token": lease,
|
||||
"idempotency_key": result_key,
|
||||
"result": {
|
||||
"status": "completed",
|
||||
"marker": "ONLINE_TASK_FLOW_OK",
|
||||
"agent_observed_task_id": task_id,
|
||||
"summary": "Agent received, claimed, heartbeated and returned the online task.",
|
||||
"artifacts": [],
|
||||
},
|
||||
},
|
||||
)
|
||||
require("result_idempotent_replay", s2, v2, "VERIFYING")
|
||||
if v2.get("idempotent") is not True:
|
||||
raise SystemExit("result_idempotency_failed")
|
||||
|
||||
s, v = call(
|
||||
args.control_base, "POST", f"/v1/tasks/{task_id}/verify", control,
|
||||
{
|
||||
"decision": "SUCCEEDED",
|
||||
"verification": {
|
||||
"marker_match": "PASS",
|
||||
"task_id_match": "PASS",
|
||||
"online_round_trip": "PASS",
|
||||
},
|
||||
},
|
||||
)
|
||||
require("verify", s, v, "SUCCEEDED")
|
||||
|
||||
s, v = call(args.control_base, "GET", f"/v1/tasks/{task_id}", control)
|
||||
require("final_control_readback", s, v, "SUCCEEDED")
|
||||
|
||||
# The executing Agent must also be able to observe the final terminal state.
|
||||
s, v = call(args.agent_base, "GET", f"/v1/tasks/{task_id}", agent)
|
||||
require("final_agent_readback", s, v, "SUCCEEDED")
|
||||
|
||||
print(json.dumps({
|
||||
"status": "ONLINE_TASK_FLOW_VERIFIED",
|
||||
"task_id": task_id,
|
||||
"final_state": v.get("status"),
|
||||
"force_push_used": False,
|
||||
"tag_rewritten": False,
|
||||
"direct_main_push": False,
|
||||
"production_changed": False,
|
||||
}, ensure_ascii=False, indent=2))
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -0,0 +1,100 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Build deterministic immutable staging artifact for AIOS Agent Task API."""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import hashlib
|
||||
import json
|
||||
from pathlib import Path
|
||||
import tarfile
|
||||
import tempfile
|
||||
|
||||
|
||||
REQUIRED = (
|
||||
"active-runtime/app/aios_agent_task_protocol.py",
|
||||
"active-runtime/app/aios_agent_task_api.py",
|
||||
"docs/gitops/openapi-agent-task-v1.yaml",
|
||||
)
|
||||
|
||||
|
||||
def sha256(path: Path) -> str:
|
||||
digest = hashlib.sha256()
|
||||
with path.open("rb") as source:
|
||||
for chunk in iter(lambda: source.read(1024 * 1024), b""):
|
||||
digest.update(chunk)
|
||||
return digest.hexdigest()
|
||||
|
||||
|
||||
def main() -> int:
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("--source-root", required=True, type=Path)
|
||||
parser.add_argument("--output-root", required=True, type=Path)
|
||||
parser.add_argument("--source-oid", required=True)
|
||||
args = parser.parse_args()
|
||||
|
||||
oid = args.source_oid.strip().lower()
|
||||
if len(oid) != 40 or any(ch not in "0123456789abcdef" for ch in oid):
|
||||
raise SystemExit("source oid must be exact 40 hex")
|
||||
source = args.source_root.resolve(strict=True)
|
||||
output = args.output_root.resolve()
|
||||
for relative in REQUIRED:
|
||||
if not (source / relative).is_file():
|
||||
raise SystemExit("required source missing: " + relative)
|
||||
|
||||
output.mkdir(parents=True, exist_ok=True)
|
||||
archive = output / f"aios-agent-task-api-{oid}.tar.gz"
|
||||
manifest = {
|
||||
"schema": "aios.agent-task-artifact.v1",
|
||||
"source_oid": oid,
|
||||
"repository_identity": "aios/aios",
|
||||
"deployment_state": "STAGING",
|
||||
"runtime_mode": "IMMUTABLE_STAGING",
|
||||
"entrypoint": "app/aios_agent_task_api.py",
|
||||
"listen": "127.0.0.1:18940",
|
||||
"plaintext_credentials_in_artifact": False,
|
||||
"included_paths": list(REQUIRED),
|
||||
}
|
||||
|
||||
with tempfile.TemporaryDirectory(prefix="agent-task-build-", dir=str(output)) as temp:
|
||||
stage = Path(temp) / "runtime"
|
||||
for relative in REQUIRED:
|
||||
src = source / relative
|
||||
if relative.startswith("active-runtime/"):
|
||||
dest = stage / Path(relative).relative_to("active-runtime")
|
||||
else:
|
||||
dest = stage / relative
|
||||
dest.parent.mkdir(parents=True, exist_ok=True)
|
||||
dest.write_bytes(src.read_bytes())
|
||||
(stage / "manifest.json").write_text(
|
||||
json.dumps(manifest, sort_keys=True, indent=2) + "\n",
|
||||
encoding="utf-8",
|
||||
)
|
||||
with tarfile.open(archive, "w:gz", format=tarfile.PAX_FORMAT) as tar:
|
||||
for path in sorted(stage.rglob("*")):
|
||||
arcname = str(path.relative_to(stage.parent)).replace("\\", "/")
|
||||
info = tar.gettarinfo(str(path), arcname=arcname)
|
||||
info.mtime = 0
|
||||
info.uid = info.gid = 0
|
||||
info.uname = info.gname = "root"
|
||||
if path.is_file():
|
||||
with path.open("rb") as fh:
|
||||
tar.addfile(info, fh)
|
||||
else:
|
||||
tar.addfile(info)
|
||||
|
||||
receipt = {
|
||||
**manifest,
|
||||
"artifact_path": str(archive),
|
||||
"artifact_sha256": sha256(archive),
|
||||
}
|
||||
receipt_path = archive.with_suffix(archive.suffix + ".receipt.json")
|
||||
receipt_path.write_text(
|
||||
json.dumps(receipt, sort_keys=True, indent=2) + "\n",
|
||||
encoding="utf-8",
|
||||
)
|
||||
print(json.dumps(receipt, sort_keys=True, indent=2))
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -45,9 +45,6 @@ class ShadowClientRBACPublicInterfaceTests(unittest.TestCase):
|
||||
identity_verifier=self.IdentityVerifier({"shadow-client"}),
|
||||
execution_enabled=False,
|
||||
identity_capabilities={"shadow-client": set(capabilities)},
|
||||
identity_actor_contexts={
|
||||
"shadow-client": {"scope_type": "system"}
|
||||
},
|
||||
)
|
||||
service = self.TaskPlaneService(store, policy=policy)
|
||||
|
||||
|
||||
@@ -1,298 +0,0 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
import os
|
||||
import sqlite3
|
||||
import sys
|
||||
import threading
|
||||
import urllib.error
|
||||
import urllib.request
|
||||
from http.server import ThreadingHTTPServer
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT = os.getcwd()
|
||||
if os.path.join(ROOT, "active-runtime") not in sys.path:
|
||||
sys.path.append(os.path.join(ROOT, "active-runtime"))
|
||||
|
||||
from app.task_plane.api import TaskPlaneRequestHandler, TaskPlaneService
|
||||
from app.task_plane.errors import ForbiddenError, NotFoundError, ValidationError
|
||||
from app.task_plane.models import TaskPlan
|
||||
from app.task_plane.policy import PolicyEngine
|
||||
from app.task_plane.store import TaskPlaneStore
|
||||
|
||||
|
||||
def _policy() -> PolicyEngine:
|
||||
return PolicyEngine(
|
||||
identity_capabilities={
|
||||
"actor-a": {"task.creator", "task.read"},
|
||||
"actor-b": {"task.creator", "task.read"},
|
||||
"reader-a": {"task.read"},
|
||||
"reader-b": {"task.read"},
|
||||
"owner": {"task.creator", "task.read", "task.read.all"},
|
||||
"unbound": {"task.creator", "task.read"},
|
||||
},
|
||||
identity_actor_contexts={
|
||||
"actor-a": {
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-a",
|
||||
"user_id": "user-a",
|
||||
"project_id": "project-a",
|
||||
"conversation_id": "conversation-a",
|
||||
"run_id": "run-a",
|
||||
},
|
||||
"reader-a": {
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-a",
|
||||
"user_id": "user-a",
|
||||
"project_id": "project-a",
|
||||
"conversation_id": "conversation-a",
|
||||
"run_id": "run-a",
|
||||
},
|
||||
"actor-b": {
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-b",
|
||||
"user_id": "user-b",
|
||||
"project_id": "project-b",
|
||||
},
|
||||
"reader-b": {
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-b",
|
||||
"user_id": "user-b",
|
||||
"project_id": "project-b",
|
||||
},
|
||||
"owner": {"scope_type": "system"},
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
def _service() -> TaskPlaneService:
|
||||
return TaskPlaneService(TaskPlaneStore(":memory:"), policy=_policy())
|
||||
|
||||
|
||||
def test_taskplan_scope_fields_round_trip_as_first_class_values() -> None:
|
||||
plan = TaskPlan.from_dict(
|
||||
{
|
||||
"task_id": "task-scope-roundtrip",
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-a",
|
||||
"user_id": "user-a",
|
||||
"project_id": "project-a",
|
||||
"conversation_id": "conversation-a",
|
||||
"run_id": "run-a",
|
||||
}
|
||||
)
|
||||
payload = plan.to_dict()
|
||||
assert TaskPlan.from_dict(payload).to_dict() == payload
|
||||
|
||||
|
||||
def test_tenant_scope_requires_tenant_id_and_system_scope_does_not() -> None:
|
||||
with pytest.raises(ValidationError) as exc:
|
||||
TaskPlan.from_dict({"task_id": "task-missing-tenant", "scope_type": "tenant"})
|
||||
assert exc.value.error_code == "TASK_SCOPE_TENANT_ID_REQUIRED"
|
||||
system = TaskPlan.from_dict(
|
||||
{"task_id": "task-system", "scope_type": "system", "run_id": "system-run"}
|
||||
)
|
||||
assert system.tenant_id is None
|
||||
|
||||
|
||||
def test_ACTOR_BINDING_AUTHZ_MERGE_uses_server_scope() -> None:
|
||||
service = _service()
|
||||
created = service.create_task(
|
||||
{"task_id": "task-a", "task_type": "code.change"},
|
||||
actor_identity="actor-a",
|
||||
)
|
||||
assert created["plan"]["tenant_id"] == "tenant-a"
|
||||
assert created["plan"]["user_id"] == "user-a"
|
||||
assert created["plan"]["project_id"] == "project-a"
|
||||
declared = service.create_task(
|
||||
{
|
||||
"task_id": "task-a-declared",
|
||||
"tenant_id": "tenant-a",
|
||||
"user_id": "user-a",
|
||||
"project_id": "project-a",
|
||||
"conversation_id": "conversation-a",
|
||||
"run_id": "run-a",
|
||||
},
|
||||
actor_identity="actor-a",
|
||||
)
|
||||
assert declared["plan"]["scope_type"] == "tenant"
|
||||
|
||||
|
||||
def test_TASK_SCOPE_MISMATCH_DENY_rejects_forged_client_scope() -> None:
|
||||
service = _service()
|
||||
with pytest.raises(ForbiddenError) as exc:
|
||||
service.create_task(
|
||||
{
|
||||
"task_id": "task-forged",
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-b",
|
||||
"user_id": "user-b",
|
||||
"project_id": "project-b",
|
||||
},
|
||||
actor_identity="actor-a",
|
||||
)
|
||||
assert exc.value.error_code == "TASK_SCOPE_MISMATCH_DENY"
|
||||
|
||||
|
||||
def test_public_create_fails_closed_without_server_actor_scope() -> None:
|
||||
service = _service()
|
||||
with pytest.raises(ForbiddenError) as exc:
|
||||
service.create_task(
|
||||
{"task_id": "task-unbound"}, actor_identity="unbound"
|
||||
)
|
||||
assert exc.value.error_code == "TASK_ACTOR_SCOPE_UNAVAILABLE"
|
||||
|
||||
|
||||
def test_TASKPLANE_LIST_AUTHZ_DENY_and_TASK_OWNERSHIP_FILTER() -> None:
|
||||
service = _service()
|
||||
service.create_task({"task_id": "task-a"}, actor_identity="actor-a")
|
||||
service.create_task({"task_id": "task-b"}, actor_identity="actor-b")
|
||||
assert [item["task_id"] for item in service.list_tasks("reader-a")] == ["task-a"]
|
||||
assert [item["task_id"] for item in service.list_tasks("reader-b")] == ["task-b"]
|
||||
|
||||
|
||||
def test_TASKPLANE_GET_AUTHZ_DENY_does_not_disclose_cross_tenant_task() -> None:
|
||||
service = _service()
|
||||
service.create_task({"task_id": "task-a"}, actor_identity="actor-a")
|
||||
with pytest.raises(NotFoundError):
|
||||
service.get_task("task-a", "reader-b")
|
||||
|
||||
|
||||
def test_http_authentication_drives_server_scope_and_cross_tenant_404() -> None:
|
||||
service = _service()
|
||||
|
||||
class Handler(TaskPlaneRequestHandler):
|
||||
pass
|
||||
|
||||
Handler.service = service
|
||||
Handler.authenticator = lambda token: {
|
||||
"token-a": "actor-a",
|
||||
"token-b": "reader-b",
|
||||
}.get(token)
|
||||
server = ThreadingHTTPServer(("127.0.0.1", 0), Handler)
|
||||
thread = threading.Thread(target=server.serve_forever, daemon=True)
|
||||
thread.start()
|
||||
base = f"http://127.0.0.1:{server.server_port}"
|
||||
try:
|
||||
request = urllib.request.Request(
|
||||
base + "/tasks",
|
||||
data=json.dumps({"task_id": "task-http-a"}).encode(),
|
||||
headers={
|
||||
"Authorization": "Bearer token-a",
|
||||
"Content-Type": "application/json",
|
||||
},
|
||||
method="POST",
|
||||
)
|
||||
with urllib.request.urlopen(request, timeout=3) as response:
|
||||
created = json.loads(response.read())
|
||||
assert created["plan"]["tenant_id"] == "tenant-a"
|
||||
|
||||
denied = urllib.request.Request(
|
||||
base + "/tasks/task-http-a",
|
||||
headers={"Authorization": "Bearer token-b"},
|
||||
)
|
||||
with pytest.raises(urllib.error.HTTPError) as exc:
|
||||
urllib.request.urlopen(denied, timeout=3)
|
||||
assert exc.value.code == 404
|
||||
finally:
|
||||
server.shutdown()
|
||||
server.server_close()
|
||||
thread.join(timeout=3)
|
||||
|
||||
|
||||
def test_mutating_endpoint_is_scoped_before_cancel() -> None:
|
||||
policy = PolicyEngine(
|
||||
execution_enabled=True,
|
||||
identity_capabilities={
|
||||
"actor-a": {"task.creator", "task.read"},
|
||||
"actor-b": {"task.cancel"},
|
||||
},
|
||||
identity_actor_contexts={
|
||||
"actor-a": {"scope_type": "tenant", "tenant_id": "tenant-a"},
|
||||
"actor-b": {"scope_type": "tenant", "tenant_id": "tenant-b"},
|
||||
},
|
||||
)
|
||||
service = TaskPlaneService(TaskPlaneStore(":memory:"), policy=policy)
|
||||
service.create_task({"task_id": "task-a"}, actor_identity="actor-a")
|
||||
with pytest.raises(NotFoundError):
|
||||
service.cancel("task-a", {"actor_identity": "actor-b"})
|
||||
|
||||
|
||||
def test_privileged_server_actor_can_cross_scope() -> None:
|
||||
service = _service()
|
||||
service.create_task({"task_id": "task-a"}, actor_identity="actor-a")
|
||||
service.create_task({"task_id": "task-b"}, actor_identity="actor-b")
|
||||
assert {item["task_id"] for item in service.list_tasks("owner")} == {
|
||||
"task-a",
|
||||
"task-b",
|
||||
}
|
||||
|
||||
|
||||
def test_scope_sql_parameters_do_not_interpolate_tenant_values() -> None:
|
||||
service = _service()
|
||||
service.create_task({"task_id": "task-a"}, actor_identity="actor-a")
|
||||
policy = PolicyEngine(
|
||||
identity_capabilities={"attacker": {"task.read"}},
|
||||
identity_actor_contexts={
|
||||
"attacker": {
|
||||
"scope_type": "tenant",
|
||||
"tenant_id": "tenant-a' OR 1=1 --",
|
||||
}
|
||||
},
|
||||
)
|
||||
attacker_service = TaskPlaneService(service.store, policy=policy)
|
||||
assert attacker_service.list_tasks("attacker") == []
|
||||
|
||||
|
||||
def test_existing_sqlite_rows_migrate_to_legacy_without_tenant_guess(tmp_path) -> None:
|
||||
path = tmp_path / "legacy.sqlite3"
|
||||
connection = sqlite3.connect(path)
|
||||
connection.execute(
|
||||
"""
|
||||
CREATE TABLE tasks (
|
||||
task_id TEXT PRIMARY KEY, revision INTEGER NOT NULL, status TEXT NOT NULL,
|
||||
task_json TEXT NOT NULL, task_type TEXT NOT NULL, root_task_id TEXT NOT NULL,
|
||||
parent_task_id TEXT, correlation_id TEXT, idempotency_key TEXT UNIQUE,
|
||||
created_at REAL NOT NULL, updated_at REAL NOT NULL, failure_reason TEXT
|
||||
)
|
||||
"""
|
||||
)
|
||||
connection.execute(
|
||||
"INSERT INTO tasks VALUES (?,?,?,?,?,?,?,?,?,?,?,?)",
|
||||
(
|
||||
"task-legacy",
|
||||
1,
|
||||
"CREATED",
|
||||
json.dumps({"task_id": "task-legacy"}),
|
||||
"code.change",
|
||||
"task-legacy",
|
||||
None,
|
||||
None,
|
||||
"legacy-key",
|
||||
1.0,
|
||||
1.0,
|
||||
None,
|
||||
),
|
||||
)
|
||||
connection.commit()
|
||||
connection.close()
|
||||
|
||||
store = TaskPlaneStore(str(path))
|
||||
task = store.get_task("task-legacy")
|
||||
assert task.task_plan.scope_type == "legacy"
|
||||
assert task.task_plan.tenant_id is None
|
||||
replay = store.create_task(
|
||||
TaskPlan.from_dict({"task_id": "task-legacy"}),
|
||||
idempotency_key="legacy-key",
|
||||
)
|
||||
assert replay.task_id == "task-legacy"
|
||||
assert tuple(
|
||||
store.conn.execute(
|
||||
"SELECT scope_type,tenant_id FROM tasks WHERE task_id='task-legacy'"
|
||||
).fetchone()
|
||||
) == ("legacy", None)
|
||||
tenant_context = _policy().actor_context("reader-a")
|
||||
with pytest.raises(NotFoundError):
|
||||
store.get_task("task-legacy", tenant_context)
|
||||
store.close()
|
||||
@@ -0,0 +1,48 @@
|
||||
from pathlib import Path
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
|
||||
|
||||
def test_model_api_exposes_governed_writer_without_legacy_global_enable():
|
||||
source = (ROOT / "active-runtime/app/aios_model_api_staging.py").read_text(encoding="utf-8")
|
||||
assert "'/createGovernedTask': ('/v2/governed-task'" in source
|
||||
assert "'/commitFilesGoverned': ('/v2/commit-files'" in source
|
||||
assert "path == '/api/governed-write-capabilities'" in source
|
||||
assert "AIOS_GITOPS_WRITES_ENABLED" not in source or "legacy_global_writes_enabled" in source
|
||||
assert "'writes_enabled': False" in source
|
||||
assert "bridge_v2_json" in source
|
||||
|
||||
|
||||
def test_model_api_routes_online_task_control_through_loopback_credential():
|
||||
source = (ROOT / "active-runtime/app/aios_model_api_staging.py").read_text(encoding="utf-8")
|
||||
assert "AIOS_AGENT_TASK_API_URL" in source
|
||||
assert "AIOS_AGENT_TASK_CONTROL_TOKEN_FILE" in source
|
||||
assert "/createOnlineAgentTask" in source
|
||||
assert "/verifyOnlineAgentTask" in source
|
||||
assert "/cancelOnlineAgentTask" in source
|
||||
assert "agent-task:control" in source
|
||||
|
||||
|
||||
def test_online_task_service_keeps_control_endpoints_off_public_nginx():
|
||||
nginx = (ROOT / "deploy/nginx/aios-agent-task-api.conf.example").read_text(encoding="utf-8")
|
||||
assert "/agent-tasks/v1/tasks/poll" in nginx
|
||||
assert "(claim|heartbeat|result)" in nginx
|
||||
assert "/v1/agents/register" not in nginx.replace(
|
||||
"# Control endpoints (/v1/agents/register, POST /v1/tasks, verify/cancel/events)", ""
|
||||
)
|
||||
|
||||
|
||||
def test_online_task_systemd_uses_runtime_credential_file():
|
||||
unit = (ROOT / "deploy/systemd/aios-agent-task-api.service").read_text(encoding="utf-8")
|
||||
assert "LoadCredential=control_token:" in unit
|
||||
assert "/run/credentials/aios-agent-task-api.service/control_token" in unit
|
||||
assert "ReadWritePaths=/var/lib/aios-agent-task" in unit
|
||||
assert "127.0.0.1" in unit
|
||||
|
||||
|
||||
def test_control_plane_snapshot_is_local_aggregate_not_missing_bridge_route():
|
||||
source = (ROOT / "active-runtime/app/aios_model_api_staging.py").read_text(encoding="utf-8")
|
||||
assert "def control_plane_snapshot(self):" in source
|
||||
assert "'snapshot_source': 'model-api-local-aggregate'" in source
|
||||
assert "return self.control_plane_snapshot()" in source
|
||||
assert "bridge_v2_json('/v2/control-plane-snapshot'" not in source
|
||||
@@ -0,0 +1,317 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from pathlib import Path
|
||||
import threading
|
||||
|
||||
import pytest
|
||||
|
||||
from aios_agent_task_protocol import TaskProtocolError, TaskStore
|
||||
|
||||
|
||||
class Clock:
|
||||
def __init__(self) -> None:
|
||||
self.value = datetime(2026, 8, 9, 12, 0, tzinfo=timezone.utc)
|
||||
|
||||
def __call__(self) -> datetime:
|
||||
return self.value
|
||||
|
||||
def advance(self, seconds: int) -> None:
|
||||
self.value += timedelta(seconds=seconds)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def env(tmp_path: Path):
|
||||
clock = Clock()
|
||||
store = TaskStore(tmp_path / "tasks.sqlite3", now=clock, token_pepper="test-pepper")
|
||||
a1 = store.register_agent(
|
||||
agent_id="openclaw-1",
|
||||
agent_type="openclaw",
|
||||
agent_version="test",
|
||||
capabilities=["shell", "git", "python"],
|
||||
)
|
||||
a2 = store.register_agent(
|
||||
agent_id="hermes-1",
|
||||
agent_type="hermes",
|
||||
agent_version="test",
|
||||
capabilities=["shell", "git", "python"],
|
||||
)
|
||||
return store, clock, a1, a2
|
||||
|
||||
|
||||
def create_task(store: TaskStore, task_id: str = "TASK-ONLINE-001"):
|
||||
return store.create_task(
|
||||
task_id=task_id,
|
||||
goal="Create a deterministic validation artifact",
|
||||
instructions="Write validation.txt containing ONLINE_TASK_OK.",
|
||||
success_criteria=[
|
||||
{"type": "file_exists", "path": "validation.txt"},
|
||||
{"type": "contains", "path": "validation.txt", "value": "ONLINE_TASK_OK"},
|
||||
],
|
||||
required_capabilities=["shell"],
|
||||
constraints={"network": "optional"},
|
||||
metadata={"suite": "online-task"},
|
||||
priority=10,
|
||||
max_attempts=3,
|
||||
idempotency_key="create:" + task_id,
|
||||
)
|
||||
|
||||
|
||||
def test_full_flow_requires_server_verification(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
create_task(store)
|
||||
|
||||
polled = store.poll_tasks(agent_id=a1["agent_id"])
|
||||
assert [item["task_id"] for item in polled] == ["TASK-ONLINE-001"]
|
||||
|
||||
claim = store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=300,
|
||||
)
|
||||
assert claim["status"] == "CLAIMED"
|
||||
assert claim["lease_token"].startswith("lease_")
|
||||
|
||||
running = store.heartbeat(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
progress={"percent": 50},
|
||||
lease_seconds=300,
|
||||
)
|
||||
assert running["status"] == "RUNNING"
|
||||
|
||||
submitted = store.submit_result(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="result:1",
|
||||
result={
|
||||
"status": "completed",
|
||||
"summary": "Created validation.txt",
|
||||
"artifacts": [{"path": "validation.txt", "sha256": "0" * 64}],
|
||||
},
|
||||
)
|
||||
assert submitted["status"] == "VERIFYING"
|
||||
assert submitted["status"] != "SUCCEEDED"
|
||||
|
||||
verified = store.verify_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
decision="SUCCEEDED",
|
||||
verification={"checks": {"content": "PASS"}},
|
||||
)
|
||||
assert verified["status"] == "SUCCEEDED"
|
||||
|
||||
|
||||
def test_atomic_claim_allows_only_one_agent(env):
|
||||
store, _clock, a1, a2 = env
|
||||
create_task(store)
|
||||
outcomes: list[tuple[str, str]] = []
|
||||
barrier = threading.Barrier(2)
|
||||
|
||||
def runner(agent_id: str) -> None:
|
||||
barrier.wait()
|
||||
try:
|
||||
store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=agent_id,
|
||||
lease_seconds=300,
|
||||
)
|
||||
outcomes.append((agent_id, "CLAIMED"))
|
||||
except TaskProtocolError as exc:
|
||||
outcomes.append((agent_id, exc.code))
|
||||
|
||||
t1 = threading.Thread(target=runner, args=(a1["agent_id"],))
|
||||
t2 = threading.Thread(target=runner, args=(a2["agent_id"],))
|
||||
t1.start()
|
||||
t2.start()
|
||||
t1.join()
|
||||
t2.join()
|
||||
|
||||
assert sum(result == "CLAIMED" for _, result in outcomes) == 1
|
||||
assert sum(result in {"TASK_NOT_CLAIMABLE", "TASK_CLAIM_CONFLICT"} for _, result in outcomes) == 1
|
||||
|
||||
|
||||
def test_expired_lease_is_requeued_and_reclaimable(env):
|
||||
store, clock, a1, a2 = env
|
||||
create_task(store)
|
||||
first = store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=30,
|
||||
)
|
||||
assert first["attempt"] == 1
|
||||
|
||||
clock.advance(31)
|
||||
available = store.poll_tasks(agent_id=a2["agent_id"])
|
||||
assert available[0]["status"] == "QUEUED"
|
||||
|
||||
second = store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a2["agent_id"],
|
||||
lease_seconds=30,
|
||||
)
|
||||
assert second["attempt"] == 2
|
||||
assert second["claimed_by"] == a2["agent_id"]
|
||||
|
||||
|
||||
def test_wrong_lease_token_is_rejected(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
create_task(store)
|
||||
store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=300,
|
||||
)
|
||||
with pytest.raises(TaskProtocolError, match="TASK_LEASE_TOKEN_INVALID"):
|
||||
store.heartbeat(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token="lease_wrong",
|
||||
progress={},
|
||||
)
|
||||
|
||||
|
||||
def test_result_submission_is_idempotent(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
create_task(store)
|
||||
claim = store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=300,
|
||||
)
|
||||
payload = {"status": "completed", "summary": "ok"}
|
||||
first = store.submit_result(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="result:same",
|
||||
result=payload,
|
||||
)
|
||||
second = store.submit_result(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="result:same",
|
||||
result=payload,
|
||||
)
|
||||
assert first["status"] == "VERIFYING"
|
||||
assert second["status"] == "VERIFYING"
|
||||
assert second["idempotent"] is True
|
||||
|
||||
with pytest.raises(TaskProtocolError, match="IDEMPOTENCY_CONFLICT"):
|
||||
store.submit_result(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="result:same",
|
||||
result={"status": "completed", "summary": "different"},
|
||||
)
|
||||
|
||||
|
||||
def test_needs_rework_returns_to_queue(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
create_task(store)
|
||||
claim = store.claim_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=300,
|
||||
)
|
||||
store.submit_result(
|
||||
task_id="TASK-ONLINE-001",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="result:1",
|
||||
result={"status": "completed"},
|
||||
)
|
||||
result = store.verify_task(
|
||||
task_id="TASK-ONLINE-001",
|
||||
decision="NEEDS_REWORK",
|
||||
verification={"reason": "success criteria not met"},
|
||||
)
|
||||
assert result["status"] == "QUEUED"
|
||||
assert result["claimed_by"] is None
|
||||
|
||||
|
||||
def test_capability_filter(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
store.create_task(
|
||||
task_id="TASK-GPU-001",
|
||||
goal="GPU task",
|
||||
instructions="Run GPU workload",
|
||||
success_criteria=[],
|
||||
required_capabilities=["gpu"],
|
||||
idempotency_key="create:gpu",
|
||||
)
|
||||
assert store.poll_tasks(agent_id=a1["agent_id"]) == []
|
||||
|
||||
|
||||
def test_create_task_idempotency_conflict(env):
|
||||
store, _clock, _a1, _a2 = env
|
||||
first = create_task(store)
|
||||
replay = create_task(store)
|
||||
assert first["task_id"] == replay["task_id"]
|
||||
assert replay["idempotent"] is True
|
||||
|
||||
with pytest.raises(TaskProtocolError, match="IDEMPOTENCY_CONFLICT"):
|
||||
store.create_task(
|
||||
task_id="TASK-ONLINE-002",
|
||||
goal="different",
|
||||
instructions="different",
|
||||
success_criteria=[],
|
||||
idempotency_key="create:TASK-ONLINE-001",
|
||||
)
|
||||
|
||||
|
||||
def test_terminal_success_retains_executing_agent_for_final_readback(env):
|
||||
store, _clock, a1, _a2 = env
|
||||
create_task(store, "TASK-TERMINAL-READBACK")
|
||||
claim = store.claim_task(
|
||||
task_id="TASK-TERMINAL-READBACK",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=120,
|
||||
)
|
||||
store.submit_result(
|
||||
task_id="TASK-TERMINAL-READBACK",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="terminal-result",
|
||||
result={"status": "completed"},
|
||||
)
|
||||
final = store.verify_task(
|
||||
task_id="TASK-TERMINAL-READBACK",
|
||||
decision="SUCCEEDED",
|
||||
verification={"criteria": "PASS"},
|
||||
)
|
||||
assert final["status"] == "SUCCEEDED"
|
||||
assert final["claimed_by"] == a1["agent_id"]
|
||||
|
||||
|
||||
def test_needs_rework_releases_agent_ownership_for_reclaim(env):
|
||||
store, _clock, a1, a2 = env
|
||||
create_task(store, "TASK-REWORK-OWNERSHIP")
|
||||
claim = store.claim_task(
|
||||
task_id="TASK-REWORK-OWNERSHIP",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_seconds=120,
|
||||
)
|
||||
store.submit_result(
|
||||
task_id="TASK-REWORK-OWNERSHIP",
|
||||
agent_id=a1["agent_id"],
|
||||
lease_token=claim["lease_token"],
|
||||
idempotency_key="rework-owner-result",
|
||||
result={"status": "completed"},
|
||||
)
|
||||
requeued = store.verify_task(
|
||||
task_id="TASK-REWORK-OWNERSHIP",
|
||||
decision="NEEDS_REWORK",
|
||||
verification={"criteria": "FAIL"},
|
||||
)
|
||||
assert requeued["status"] == "QUEUED"
|
||||
assert requeued["claimed_by"] is None
|
||||
reclaimed = store.claim_task(
|
||||
task_id="TASK-REWORK-OWNERSHIP",
|
||||
agent_id=a2["agent_id"],
|
||||
lease_seconds=120,
|
||||
)
|
||||
assert reclaimed["claimed_by"] == a2["agent_id"]
|
||||
@@ -2,7 +2,6 @@ import base64
|
||||
from pathlib import Path
|
||||
|
||||
from aios_authority_git_executor import AuthorityGitExecutor
|
||||
from aios_chatgpt_bridge_v2 import BridgeV2Adapter, OPERATION_REQUIRED_FIELDS
|
||||
from aios_chatgpt_bridge_v2_server import BridgeHandler, RuntimeConfig
|
||||
from aios_continuation_executor import GiteaContinuationRemote, ServerSideCredentialProjection
|
||||
from aios_continuation_lease import WriterLeaseRequestV2
|
||||
@@ -27,7 +26,6 @@ class Probe:
|
||||
|
||||
def test_create_new_lease_is_exact_base_and_absent():
|
||||
request = WriterLeaseRequestV2(
|
||||
job_id="job-bridge-runtime-lineage-test",
|
||||
task_id="TASK-P0-BRIDGE-V21-DURABLE-RECONCILIATION-20260809",
|
||||
repository="aios/aios",
|
||||
branch_mode="CREATE_NEW",
|
||||
@@ -41,7 +39,6 @@ def test_create_new_lease_is_exact_base_and_absent():
|
||||
|
||||
def test_create_new_lease_rejects_existing_target():
|
||||
request = WriterLeaseRequestV2(
|
||||
job_id="job-bridge-runtime-lineage-test",
|
||||
task_id="task",
|
||||
repository="aios/aios",
|
||||
branch_mode="CREATE_NEW",
|
||||
@@ -78,7 +75,7 @@ def test_slash_branch_uses_paginated_collection():
|
||||
def test_http_route_contract_is_complete():
|
||||
expected = {
|
||||
"/v2/runtime", "/v2/capabilities", "/v2/governed-task",
|
||||
"/v2/writer-lease", "/v2/writer-lease/bootstrap", "/v2/writer-lease/renew", "/v2/commit-file", "/v2/commit-files",
|
||||
"/v2/writer-lease", "/v2/commit-file", "/v2/commit-files",
|
||||
"/v2/job", "/v2/job/resume", "/v2/job/reconcile",
|
||||
"/v2/tag/approval-request",
|
||||
"/v2/tag/publish",
|
||||
@@ -91,121 +88,3 @@ def test_http_route_contract_is_complete():
|
||||
from aios_chatgpt_bridge_v2_server import ROUTES
|
||||
|
||||
assert set(ROUTES) == expected
|
||||
|
||||
|
||||
def test_renew_writer_lease_is_exposed_in_action_and_capability_contracts():
|
||||
adapter = BridgeV2Adapter(runtime_status={})
|
||||
capabilities = adapter.getGovernedWriteCapabilities()
|
||||
contract = capabilities["operation_contracts"]["renewWriterLease"]
|
||||
|
||||
assert contract["method"] == "POST"
|
||||
assert contract["path"] == "/v2/writer-lease/renew"
|
||||
assert tuple(contract["required_fields"]) == OPERATION_REQUIRED_FIELDS["renewWriterLease"]
|
||||
assert "runtime_rehydrated" in contract["response_fields"]
|
||||
|
||||
action_schema = Path("docs/gitops/openapi-chatgpt-actions.yaml").read_text(encoding="utf-8")
|
||||
assert "/v2/writer-lease/renew:" in action_schema
|
||||
assert "operationId: renewWriterLease" in action_schema
|
||||
assert "BridgeWriterLeaseRecoveryRequest" in action_schema
|
||||
assert "BridgeWriterLeaseRecoveryResponse" in action_schema
|
||||
assert {
|
||||
capabilities["operation_contracts"][operation]["path"]
|
||||
for operation in (
|
||||
"acquireWriterLease",
|
||||
"renewWriterLease",
|
||||
"commitFileGoverned",
|
||||
"commitFilesGoverned",
|
||||
)
|
||||
} == {
|
||||
"/v2/writer-lease",
|
||||
"/v2/writer-lease/renew",
|
||||
"/v2/commit-file",
|
||||
"/v2/commit-files",
|
||||
}
|
||||
|
||||
|
||||
def test_bootstrap_writer_lease_is_exposed_in_action_and_capability_contracts():
|
||||
adapter = BridgeV2Adapter(runtime_status={})
|
||||
capabilities = adapter.getGovernedWriteCapabilities()
|
||||
contract = capabilities["operation_contracts"]["bootstrapWriterLease"]
|
||||
|
||||
assert contract["method"] == "POST"
|
||||
assert contract["path"] == "/v2/writer-lease/bootstrap"
|
||||
assert tuple(contract["required_fields"]) == OPERATION_REQUIRED_FIELDS["bootstrapWriterLease"]
|
||||
assert "bootstrapped" in contract["response_fields"]
|
||||
assert "durable_branch_lease_present" in contract["response_fields"]
|
||||
|
||||
action_schema = Path("docs/gitops/openapi-chatgpt-actions.yaml").read_text(encoding="utf-8")
|
||||
assert "/v2/writer-lease/bootstrap:" in action_schema
|
||||
assert "operationId: bootstrapWriterLease" in action_schema
|
||||
assert "bootstrapped: {type: boolean}" in action_schema
|
||||
assert "durable_branch_lease_present: {type: boolean}" in action_schema
|
||||
|
||||
|
||||
|
||||
def test_http_allows_explicit_job_id_on_lease_and_commit_requests():
|
||||
handler = BridgeHandler.__new__(BridgeHandler)
|
||||
lease_body = {
|
||||
"schema": WriterLeaseRequestV2.schema,
|
||||
"task_id": "task",
|
||||
"job_id": "job-explicit-runtime-lineage",
|
||||
"repository": "aios/aios",
|
||||
"branch_mode": "CONTINUE_EXISTING",
|
||||
"base_oid": "a" * 40,
|
||||
"expected_branch_oid": "a" * 40,
|
||||
"target_branch": "agent/explicit-runtime-lineage",
|
||||
"generation": 2,
|
||||
"lease_seconds": 120,
|
||||
}
|
||||
handler._allowed("acquireWriterLease", lease_body)
|
||||
handler._allowed(
|
||||
"commitFilesGoverned",
|
||||
{
|
||||
"task_id": "task",
|
||||
"job_id": "job-explicit-runtime-lineage",
|
||||
"repository": "aios/aios",
|
||||
"expected_branch_oid": "a" * 40,
|
||||
"target_branch": "agent/explicit-runtime-lineage",
|
||||
"generation": 2,
|
||||
"idempotency_key": "commit-explicit-runtime-lineage",
|
||||
"files": {"docs/x.txt": "eA=="},
|
||||
"message": "test",
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
def test_http_allows_omitted_job_id_for_durable_server_resolution():
|
||||
handler = BridgeHandler.__new__(BridgeHandler)
|
||||
lease_body = {
|
||||
"schema": WriterLeaseRequestV2.schema,
|
||||
"task_id": "task",
|
||||
"repository": "aios/aios",
|
||||
"branch_mode": "CONTINUE_EXISTING",
|
||||
"base_oid": "a" * 40,
|
||||
"expected_branch_oid": "a" * 40,
|
||||
"target_branch": "agent/implicit-runtime-lineage",
|
||||
"generation": 2,
|
||||
"lease_seconds": 120,
|
||||
}
|
||||
handler._allowed("acquireWriterLease", lease_body)
|
||||
handler._allowed(
|
||||
"commitFilesGoverned",
|
||||
{
|
||||
"task_id": "task",
|
||||
"repository": "aios/aios",
|
||||
"expected_branch_oid": "a" * 40,
|
||||
"target_branch": "agent/implicit-runtime-lineage",
|
||||
"generation": 2,
|
||||
"idempotency_key": "commit-implicit-runtime-lineage",
|
||||
"files": {"docs/x.txt": "eA=="},
|
||||
"message": "test",
|
||||
},
|
||||
)
|
||||
|
||||
|
||||
def test_capability_contracts_return_exact_job_id_lineage():
|
||||
contracts = BridgeV2Adapter(runtime_status={}).getGovernedWriteCapabilities()["operation_contracts"]
|
||||
for operation in ("acquireWriterLease", "bootstrapWriterLease", "renewWriterLease"):
|
||||
assert "job_id" in contracts[operation]["response_fields"]
|
||||
for operation in ("commitFileGoverned", "commitFilesGoverned"):
|
||||
assert "job_id" in contracts[operation]["response_fields"]
|
||||
|
||||
@@ -0,0 +1,192 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime, timedelta, timezone
|
||||
import importlib.util
|
||||
import json
|
||||
from pathlib import Path
|
||||
import tempfile
|
||||
import unittest
|
||||
from unittest.mock import patch
|
||||
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
APP = ROOT / "active-runtime" / "app"
|
||||
SPEC = importlib.util.spec_from_file_location(
|
||||
"aios_durable_owner_authorization",
|
||||
APP / "aios_durable_owner_authorization.py",
|
||||
)
|
||||
BROKER_MODULE = importlib.util.module_from_spec(SPEC)
|
||||
import sys
|
||||
sys.modules[SPEC.name] = BROKER_MODULE
|
||||
assert SPEC.loader is not None
|
||||
SPEC.loader.exec_module(BROKER_MODULE)
|
||||
|
||||
AuthorizationError = BROKER_MODULE.AuthorizationError
|
||||
Broker = BROKER_MODULE.Broker
|
||||
Constraints = BROKER_MODULE.Constraints
|
||||
DurablePolicy = BROKER_MODULE.DurablePolicy
|
||||
ReceiptStore = BROKER_MODULE.ReceiptStore
|
||||
Upstream = BROKER_MODULE.Upstream
|
||||
|
||||
CALLER = "chatgpt-governed-implementer-singlehost"
|
||||
SCOPE = "release:candidate"
|
||||
SOURCE = "257c7a8df433f75b364bcaf69cbbc32e8c2e5d38"
|
||||
|
||||
|
||||
def policy(*, expires_at=None, capabilities=None):
|
||||
return DurablePolicy(
|
||||
grant_id="grant-caller-projection-v1",
|
||||
owner_identity="owner-durable-policy",
|
||||
account=CALLER,
|
||||
capabilities=frozenset(capabilities or {SCOPE}),
|
||||
issued_at=datetime.now(timezone.utc) - timedelta(minutes=1),
|
||||
not_before=datetime.now(timezone.utc) - timedelta(minutes=1),
|
||||
persistent_until_revoked=True,
|
||||
constraints=Constraints(
|
||||
repository="aios/aios",
|
||||
account=CALLER,
|
||||
allowed_target_prefix="agent/",
|
||||
release_version_prefix="V2R2C",
|
||||
max_task_risk="engineering",
|
||||
),
|
||||
policy_sha256="a" * 64,
|
||||
expires_at=expires_at,
|
||||
)
|
||||
|
||||
|
||||
def make_broker(root):
|
||||
return Broker(
|
||||
policy_path=root / "policy.json",
|
||||
revocation_path=root / "revocations.json",
|
||||
store=ReceiptStore(root / "receipts.sqlite3"),
|
||||
upstream=Upstream("http://127.0.0.1:1", "not-used"),
|
||||
expected_account=CALLER,
|
||||
expected_peer_user="test-peer",
|
||||
allowed_peer_users=frozenset({"test-peer"}),
|
||||
)
|
||||
|
||||
|
||||
class CallerProjectionTests(unittest.TestCase):
|
||||
def project(self, broker, request=None):
|
||||
return broker.project_caller(
|
||||
CALLER,
|
||||
SCOPE,
|
||||
request or {"source_commit": SOURCE, "version": "V2R2C2"},
|
||||
idempotency_key="caller-projection-test-v1",
|
||||
)
|
||||
|
||||
def test_positive_projection_ack_and_effective_scope(self):
|
||||
broker = make_broker(Path(tempfile.mkdtemp()))
|
||||
with patch.object(broker, "policy", return_value=policy()):
|
||||
value = self.project(broker)
|
||||
credential = value.pop("_credential")
|
||||
self.assertGreaterEqual(len(credential), 32)
|
||||
self.assertEqual("ACKED", value["projection_state"])
|
||||
self.assertEqual("REHYDRATED", value["credential_state"])
|
||||
self.assertEqual([SCOPE], value["effective_scopes"])
|
||||
self.assertFalse(value["secret_exposed"])
|
||||
|
||||
def test_secret_material_is_not_durable(self):
|
||||
root = Path(tempfile.mkdtemp())
|
||||
broker = make_broker(root)
|
||||
with patch.object(broker, "policy", return_value=policy()):
|
||||
value = self.project(broker)
|
||||
credential = value.pop("_credential")
|
||||
durable = (root / "receipts.sqlite3").read_bytes()
|
||||
self.assertNotIn(credential.encode(), durable)
|
||||
self.assertNotIn("credential_material", durable.decode("utf-8", "ignore"))
|
||||
self.assertFalse(value["secret_exposed"])
|
||||
|
||||
def test_caller_mismatch_is_denied(self):
|
||||
broker = make_broker(Path(tempfile.mkdtemp()))
|
||||
with self.assertRaises(AuthorizationError) as caught:
|
||||
broker.project_caller(
|
||||
"other-caller",
|
||||
SCOPE,
|
||||
{},
|
||||
idempotency_key="caller-projection-test-v1",
|
||||
)
|
||||
self.assertEqual("CALLER_MISMATCH", caught.exception.code)
|
||||
|
||||
def test_scope_mismatch_is_denied(self):
|
||||
broker = make_broker(Path(tempfile.mkdtemp()))
|
||||
with self.assertRaises(AuthorizationError) as caught:
|
||||
broker.project_caller(
|
||||
CALLER,
|
||||
"release:deploy",
|
||||
{},
|
||||
idempotency_key="caller-projection-test-v1",
|
||||
)
|
||||
self.assertEqual("SCOPE_MISMATCH", caught.exception.code)
|
||||
|
||||
def test_policy_missing_revoked_and_expired_fail_closed(self):
|
||||
broker = make_broker(Path(tempfile.mkdtemp()))
|
||||
for code in ("POLICY_FILE_INVALID", "OWNER_GRANT_REVOKED", "POLICY_EXPIRED"):
|
||||
with self.subTest(code=code):
|
||||
with patch.object(
|
||||
broker,
|
||||
"policy",
|
||||
side_effect=AuthorizationError(code, 403),
|
||||
):
|
||||
with self.assertRaises(AuthorizationError) as caught:
|
||||
self.project(broker)
|
||||
self.assertEqual(code, caught.exception.code)
|
||||
|
||||
def test_invalid_projection_state_fails_closed(self):
|
||||
with self.assertRaises(AuthorizationError) as caught:
|
||||
Broker.validate_caller_projection(
|
||||
{
|
||||
"caller": CALLER,
|
||||
"scope": SCOPE,
|
||||
"projection_state": "CORRUPT",
|
||||
"credential_state": "REHYDRATED",
|
||||
"rehydration_state": "RECOVERED",
|
||||
"effective_scopes": [SCOPE],
|
||||
"secret_exposed": False,
|
||||
}
|
||||
)
|
||||
self.assertEqual("PROJECTION_STATE_INVALID", caught.exception.code)
|
||||
|
||||
def test_restart_rehydrates_from_durable_metadata(self):
|
||||
root = Path(tempfile.mkdtemp())
|
||||
first = make_broker(root)
|
||||
with patch.object(first, "policy", return_value=policy()):
|
||||
initial = self.project(first)
|
||||
first_credential = initial.pop("_credential")
|
||||
second = make_broker(root)
|
||||
with patch.object(second, "policy", return_value=policy()):
|
||||
recovered = self.project(second)
|
||||
second_credential = recovered.pop("_credential")
|
||||
self.assertNotEqual(first_credential, second_credential)
|
||||
self.assertEqual("RECOVERED", recovered["rehydration_state"])
|
||||
self.assertEqual("ACKED", recovered["projection_state"])
|
||||
|
||||
def test_expired_policy_loader_rejects(self):
|
||||
root = Path(tempfile.mkdtemp())
|
||||
policy_path = root / "policy.json"
|
||||
payload = {
|
||||
"schema": "aios.owner-durable-authorization.v1",
|
||||
"grant_id": "grant-caller-projection-v1",
|
||||
"owner_identity": "owner-durable-policy",
|
||||
"account": CALLER,
|
||||
"capabilities": [SCOPE],
|
||||
"issued_at": "2026-08-22T00:00:00Z",
|
||||
"not_before": "2026-08-22T00:00:00Z",
|
||||
"expires_at": "2026-08-22T01:00:00Z",
|
||||
"persistent_until_revoked": True,
|
||||
"constraints": {
|
||||
"repository": "aios/aios",
|
||||
"allowed_target_prefix": "agent/",
|
||||
"release_version_prefix": "V2R2C",
|
||||
"max_task_risk": "engineering",
|
||||
},
|
||||
}
|
||||
policy_path.write_text(json.dumps(payload), encoding="utf-8")
|
||||
with patch.object(BROKER_MODULE, "validate_root_file", lambda path, **kwargs: path):
|
||||
with self.assertRaises(AuthorizationError) as caught:
|
||||
DurablePolicy.load(policy_path)
|
||||
self.assertEqual("POLICY_EXPIRED", caught.exception.code)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
@@ -0,0 +1,191 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
import sys
|
||||
import types
|
||||
from pathlib import Path
|
||||
from types import SimpleNamespace
|
||||
from unittest.mock import patch
|
||||
|
||||
import pytest
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
APP = ROOT / "active-runtime" / "app"
|
||||
if str(APP) not in sys.path:
|
||||
sys.path.insert(0, str(APP))
|
||||
|
||||
try:
|
||||
import orchestrator.action_monitor
|
||||
except ModuleNotFoundError:
|
||||
_orchestrator = types.ModuleType("orchestrator")
|
||||
_action_monitor = types.ModuleType("orchestrator.action_monitor")
|
||||
_action_monitor.ActionMonitor = type("ActionMonitor", (), {})
|
||||
_orchestrator.__path__ = []
|
||||
sys.modules["orchestrator"] = _orchestrator
|
||||
sys.modules["orchestrator.action_monitor"] = _action_monitor
|
||||
|
||||
import aios_durable_owner_authorization as broker_api
|
||||
import aios_model_api_staging as model_api
|
||||
|
||||
|
||||
CALLER = "chatgpt-governed-implementer-singlehost"
|
||||
SCOPE = "release:candidate"
|
||||
|
||||
|
||||
def _invoke_capabilities(
|
||||
headers,
|
||||
*,
|
||||
auth_result=None,
|
||||
auth_side_effect=None,
|
||||
projection_client=None,
|
||||
):
|
||||
handler = object.__new__(model_api.H)
|
||||
handler.headers = headers
|
||||
responses = []
|
||||
handler.sendj = lambda value, code=200: responses.append((value, code))
|
||||
if auth_side_effect is not None:
|
||||
auth_patch = patch.object(
|
||||
model_api, "authenticate", side_effect=auth_side_effect
|
||||
)
|
||||
else:
|
||||
auth_patch = patch.object(
|
||||
model_api, "authenticate", return_value=auth_result
|
||||
)
|
||||
client = (
|
||||
projection_client
|
||||
if projection_client is not None
|
||||
else model_api.CALLER_PROJECTION_CLIENT
|
||||
)
|
||||
with auth_patch, patch.object(model_api, "CALLER_PROJECTION_CLIENT", client):
|
||||
model_api.H.caller_capabilities(handler)
|
||||
assert len(responses) == 1
|
||||
return responses[0]
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"headers",
|
||||
[
|
||||
{},
|
||||
{"Authorization": "Bearer invalid"},
|
||||
{"Authorization": "Bearer unknown"},
|
||||
{"X-AIOS-Token": "rec-none"},
|
||||
],
|
||||
)
|
||||
def test_anonymous_invalid_unknown_and_rec_none_are_structured_401(headers):
|
||||
payload, status = _invoke_capabilities(headers, auth_result=None)
|
||||
assert status == 401
|
||||
assert payload == {"error": "unauthorized", "hint": "Bearer token required"}
|
||||
assert "Traceback" not in json.dumps(payload)
|
||||
|
||||
|
||||
def test_authenticate_exception_is_fail_closed_without_socket_abort():
|
||||
payload, status = _invoke_capabilities(
|
||||
{"Authorization": "Bearer malformed"},
|
||||
auth_side_effect=RuntimeError("malformed"),
|
||||
)
|
||||
assert status == 401
|
||||
assert payload["error"] == "unauthorized"
|
||||
assert "Traceback" not in json.dumps(payload)
|
||||
|
||||
|
||||
def test_authorized_capability_introspection_remains_secret_free():
|
||||
row = {
|
||||
"account": CALLER,
|
||||
"role": "user",
|
||||
"scopes": ["agent-task:control", "branch:create", "file:write"],
|
||||
}
|
||||
result = SimpleNamespace(
|
||||
public={
|
||||
"projection_state": "ACKED",
|
||||
"credential_state": "ACTIVE",
|
||||
"rehydration_state": "REHYDRATED",
|
||||
"effective_scopes": [SCOPE],
|
||||
},
|
||||
credential_loaded=True,
|
||||
)
|
||||
|
||||
class _ProjectionClient:
|
||||
def rehydrate(self, caller, scope, request):
|
||||
assert caller == CALLER
|
||||
assert scope == SCOPE
|
||||
assert request == {}
|
||||
return result
|
||||
|
||||
payload, status = _invoke_capabilities(
|
||||
{"Authorization": "Bearer valid"},
|
||||
auth_result=row,
|
||||
projection_client=_ProjectionClient(),
|
||||
)
|
||||
assert status == 200
|
||||
assert payload["caller_identity"] == CALLER
|
||||
assert payload["admitted"] is True
|
||||
assert SCOPE in payload["effective_scopes"]
|
||||
assert payload["projection_state"] == "ACKED"
|
||||
assert payload["credential_state"] == "ACTIVE"
|
||||
assert payload["rehydration_state"] == "REHYDRATED"
|
||||
assert payload["secret_exposed"] is False
|
||||
assert "_credential" not in json.dumps(payload)
|
||||
|
||||
|
||||
class _FakeBroker:
|
||||
def __init__(self):
|
||||
self.calls = []
|
||||
|
||||
def project_caller(self, caller, scope, request, *, idempotency_key):
|
||||
self.calls.append((caller, scope, request, idempotency_key))
|
||||
return {
|
||||
"caller": caller,
|
||||
"scope": scope,
|
||||
"projection_state": "ACKED",
|
||||
"credential_state": "ACTIVE",
|
||||
"rehydration_state": "REHYDRATED",
|
||||
"effective_scopes": [SCOPE],
|
||||
"_credential": "x" * 40,
|
||||
}
|
||||
|
||||
|
||||
def _dispatch_broker(path="/v1/project-caller", *, peer=True, account=True):
|
||||
broker = _FakeBroker()
|
||||
handler = object.__new__(broker_api.Handler)
|
||||
handler.path = path
|
||||
handler.broker = broker
|
||||
handler.peer_authorized = lambda: peer
|
||||
handler.account_authorized = lambda: account
|
||||
handler.body = lambda: {
|
||||
"caller": CALLER,
|
||||
"scope": SCOPE,
|
||||
"idempotency_key": "runtime-regression-test-20260823",
|
||||
"request": {
|
||||
"repository": "aios/aios",
|
||||
"target_branch": "agent/runtime-regression-test",
|
||||
"release_version": "V2R2C2",
|
||||
},
|
||||
}
|
||||
responses = []
|
||||
handler.send_json = lambda value, status, internal_credential=None: responses.append(
|
||||
(value, status, internal_credential)
|
||||
)
|
||||
broker_api.Handler.do_POST(handler)
|
||||
return responses, broker
|
||||
|
||||
|
||||
def test_existing_project_caller_route_reaches_handler_without_public_secret():
|
||||
responses, broker = _dispatch_broker()
|
||||
assert len(responses) == 1
|
||||
value, status, internal_credential = responses[0]
|
||||
assert status == 200
|
||||
assert broker.calls
|
||||
assert broker.calls[0][0:2] == (CALLER, SCOPE)
|
||||
assert internal_credential == "x" * 40
|
||||
assert "_credential" not in value
|
||||
assert "x" * 40 not in json.dumps(value)
|
||||
|
||||
|
||||
def test_unauthorized_peer_is_denied_and_unregistered_route_is_not_added():
|
||||
responses, broker = _dispatch_broker(peer=False)
|
||||
assert responses[0][1] == 403
|
||||
assert broker.calls == []
|
||||
|
||||
responses, broker = _dispatch_broker(path="/v1/not-registered")
|
||||
assert responses[0][1] == 404
|
||||
assert broker.calls == []
|
||||
@@ -1,304 +0,0 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from pathlib import Path
|
||||
import re
|
||||
import json
|
||||
from types import SimpleNamespace
|
||||
|
||||
import pytest
|
||||
import yaml
|
||||
|
||||
from aios_chatgpt_bridge_v2_server import GovernedRuntime
|
||||
from aios_continuation_executor import ContinuationExecutor
|
||||
from aios_continuation_lease import WriterLeaseRequestV2
|
||||
from aios_gitops_job_controller import JobState
|
||||
|
||||
|
||||
TASK = "continuation-normalization-test"
|
||||
JOB = "job-continuation-normalization-test"
|
||||
REPO = "aios/aios"
|
||||
BASE = "a" * 40
|
||||
BRANCH = "agent/continuation-normalization-test"
|
||||
|
||||
|
||||
class Authority:
|
||||
def __init__(self, head: str | None) -> None:
|
||||
self.head = head
|
||||
|
||||
def read_ref(self, ref: str) -> str | None:
|
||||
assert ref == "refs/heads/" + BRANCH
|
||||
return self.head
|
||||
|
||||
def commit_exists(self, oid: str) -> bool:
|
||||
return oid == BASE
|
||||
|
||||
def branch_exists(self, branch: str) -> bool:
|
||||
assert branch == BRANCH
|
||||
return self.head is not None
|
||||
|
||||
def branch_head(self, branch: str) -> str:
|
||||
assert branch == BRANCH and self.head is not None
|
||||
return self.head
|
||||
|
||||
def is_ancestor(self, older: str, newer: str) -> bool:
|
||||
return older == BASE and newer == BASE
|
||||
|
||||
|
||||
def lease(**overrides: object) -> WriterLeaseRequestV2:
|
||||
values = {
|
||||
"task_id": TASK,
|
||||
"job_id": JOB,
|
||||
"repository": REPO,
|
||||
"branch_mode": "CREATE_NEW",
|
||||
"base_oid": BASE,
|
||||
"expected_branch_oid": BASE,
|
||||
"target_branch": BRANCH,
|
||||
"generation": 1,
|
||||
"lease_seconds": 120,
|
||||
}
|
||||
values.update(overrides)
|
||||
return WriterLeaseRequestV2(**values)
|
||||
|
||||
|
||||
def runtime(*, head: str | None = BASE, **binding_overrides: object) -> GovernedRuntime:
|
||||
value = GovernedRuntime.__new__(GovernedRuntime)
|
||||
value.authority = Authority(head)
|
||||
binding = {
|
||||
"task_id": TASK,
|
||||
"job_id": JOB,
|
||||
"repository": REPO,
|
||||
"base_oid": BASE,
|
||||
"expected_branch_oid": BASE,
|
||||
"target_branch": BRANCH,
|
||||
"generation": 1,
|
||||
"branch_mode": "CREATE_NEW",
|
||||
}
|
||||
binding.update(binding_overrides)
|
||||
value.bindings = {TASK: binding}
|
||||
return value
|
||||
|
||||
|
||||
def test_same_lineage_create_new_normalizes_to_continue_existing() -> None:
|
||||
normalized = runtime()._normalize_runtime_lease(lease())
|
||||
assert normalized.branch_mode == "CONTINUE_EXISTING"
|
||||
assert normalized.task_id == TASK
|
||||
assert normalized.job_id == JOB
|
||||
assert normalized.repository == REPO
|
||||
assert normalized.target_branch == BRANCH
|
||||
assert normalized.expected_branch_oid == BASE
|
||||
assert normalized.generation == 1
|
||||
|
||||
|
||||
def test_unprovisioned_create_new_stays_create_new() -> None:
|
||||
assert runtime(head=None)._normalize_runtime_lease(lease()).branch_mode == "CREATE_NEW"
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
("field", "value"),
|
||||
(
|
||||
("repository", "other/repo"),
|
||||
("base_oid", "b" * 40),
|
||||
("expected_branch_oid", "b" * 40),
|
||||
("target_branch", "agent/foreign"),
|
||||
("generation", 2),
|
||||
),
|
||||
)
|
||||
def test_binding_mismatch_fails_closed(field: str, value: object) -> None:
|
||||
with pytest.raises(RuntimeError, match="LEASE_BINDING_CONFLICT"):
|
||||
runtime(**{field: value})._normalize_runtime_lease(lease())
|
||||
|
||||
|
||||
def test_drifted_existing_head_fails_closed() -> None:
|
||||
with pytest.raises(RuntimeError, match="EXPECTED_BRANCH_OID_CONFLICT"):
|
||||
runtime(head="b" * 40)._normalize_runtime_lease(lease())
|
||||
|
||||
|
||||
def test_openapi_contract_parses_and_requires_exact_commit_lineage() -> None:
|
||||
document = yaml.safe_load(
|
||||
Path("docs/gitops/openapi-chatgpt-actions.yaml").read_text(encoding="utf-8")
|
||||
)
|
||||
schema = document["components"]["schemas"]["BridgeCommitResponse"]
|
||||
assert {
|
||||
"job_id",
|
||||
"commit_oid",
|
||||
"commit_visible",
|
||||
"branch_mode",
|
||||
"target_branch",
|
||||
"generation",
|
||||
} <= set(schema["required"])
|
||||
|
||||
|
||||
def test_changed_production_files_pass_secret_scan() -> None:
|
||||
paths = [
|
||||
Path("active-runtime/app/aios_authority_git_executor.py"),
|
||||
Path("active-runtime/app/aios_chatgpt_bridge_v2_server.py"),
|
||||
Path("active-runtime/app/aios_continuation_executor.py"),
|
||||
Path("active-runtime/app/aios_continuation_lease.py"),
|
||||
Path("active-runtime/app/authorized_task_writer_provisioning.py"),
|
||||
Path("active-runtime/app/aios_chatgpt_bridge_v2.py"),
|
||||
Path("docs/gitops/openapi-chatgpt-actions.yaml"),
|
||||
]
|
||||
expression = re.compile(
|
||||
r"(?i)(BEGIN (RSA|OPENSSH|EC) PRIVATE KEY|"
|
||||
r"authorization:\\s*bearer\\s+[A-Za-z0-9._-]{12,}|"
|
||||
r"api[_-]?key\\s*[:=]\\s*['\\\"][^'\\\"]{12,})"
|
||||
)
|
||||
hits = []
|
||||
for path in paths:
|
||||
for number, line in enumerate(
|
||||
path.read_text(encoding="utf-8", errors="replace").splitlines(), 1
|
||||
):
|
||||
if expression.search(line):
|
||||
hits.append(f"{path}:{number}")
|
||||
assert hits == []
|
||||
|
||||
|
||||
def test_remote_verified_job_is_not_completed_when_head_cas_callback_fails() -> None:
|
||||
class Store:
|
||||
def read(self, _job_id):
|
||||
return {
|
||||
"branch": BRANCH,
|
||||
"request_json": json.dumps(
|
||||
{"binding": {"expected_branch_oid": BASE}}
|
||||
),
|
||||
}
|
||||
|
||||
class Controller:
|
||||
def __init__(self):
|
||||
self.store = Store()
|
||||
self.states = []
|
||||
|
||||
def advance(self, _job_id, state, **_kwargs):
|
||||
self.states.append(state)
|
||||
return SimpleNamespace(as_dict=lambda: {"state": state.value})
|
||||
|
||||
executor = ContinuationExecutor.__new__(ContinuationExecutor)
|
||||
executor.controller = Controller()
|
||||
executor._observe_remote = lambda _branch, oid: (oid, oid, True)
|
||||
|
||||
def fail_cas(_oid):
|
||||
raise RuntimeError("LEASE_CAS_CONFLICT")
|
||||
|
||||
with pytest.raises(RuntimeError, match="LEASE_CAS_CONFLICT"):
|
||||
executor._finish_remote_verification(
|
||||
JOB,
|
||||
"b" * 40,
|
||||
"b" * 40,
|
||||
before_complete=fail_cas,
|
||||
)
|
||||
|
||||
assert executor.controller.states == [JobState.REMOTE_VERIFIED]
|
||||
|
||||
|
||||
@pytest.mark.parametrize("branch_mode", ("CREATE_NEW", "CONTINUE_EXISTING"))
|
||||
def test_executor_preflight_accepts_the_exact_mode_dependent_normalized_shape(
|
||||
branch_mode: str,
|
||||
) -> None:
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from aios_continuation_executor import (
|
||||
ContinuationPermissions,
|
||||
ContinuationRuntimeBinding,
|
||||
)
|
||||
|
||||
request = lease(branch_mode=branch_mode)
|
||||
now = datetime.now(timezone.utc)
|
||||
executor = ContinuationExecutor.__new__(ContinuationExecutor)
|
||||
executor.authority = Authority(None if branch_mode == "CREATE_NEW" else BASE)
|
||||
executor.now = lambda: now
|
||||
executor._max_generation = lambda _branch, *, exclude_job_id=None: 0
|
||||
executor.workspace = SimpleNamespace(
|
||||
is_detached=lambda: True,
|
||||
head_oid=lambda: BASE,
|
||||
)
|
||||
executor.remote = SimpleNamespace(repository_identity=lambda: REPO)
|
||||
binding = ContinuationRuntimeBinding(
|
||||
request=request,
|
||||
job_id=JOB,
|
||||
workspace_path=Path("/tmp/continuation-normalization-test"),
|
||||
lease_not_before=now - timedelta(seconds=1),
|
||||
lease_expires_at=now + timedelta(seconds=60),
|
||||
original_receipt_valid=True,
|
||||
original_grant_consumed=True,
|
||||
permissions=ContinuationPermissions(),
|
||||
)
|
||||
|
||||
executor._validate_preflight(binding, job_id=JOB)
|
||||
|
||||
|
||||
def test_create_new_executor_resumes_executing_and_uses_absence_bound_writers() -> None:
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from aios_continuation_executor import (
|
||||
ContinuationCommit,
|
||||
ContinuationPermissions,
|
||||
ContinuationRuntimeBinding,
|
||||
)
|
||||
|
||||
new_oid = "b" * 40
|
||||
|
||||
class Store:
|
||||
def read(self, _job_id: str) -> dict[str, object]:
|
||||
return {"state": "EXECUTING", "revision": 4}
|
||||
|
||||
class Controller:
|
||||
def __init__(self) -> None:
|
||||
self.store = Store()
|
||||
self.advanced: list[str] = []
|
||||
|
||||
def advance(self, _job_id: str, state: JobState, **_kwargs: object) -> object:
|
||||
self.advanced.append(state.value)
|
||||
return SimpleNamespace(as_dict=lambda: {"state": state.value})
|
||||
|
||||
class Workspace:
|
||||
def commit_files(self, _files: object, _message: str) -> str:
|
||||
return new_oid
|
||||
|
||||
class CreateAuthority:
|
||||
def __init__(self) -> None:
|
||||
self.created: list[tuple[str, str, str]] = []
|
||||
|
||||
def is_ancestor(self, older: str, newer: str) -> bool:
|
||||
return older == BASE and newer == new_oid
|
||||
|
||||
def create_task_branch_cas(self, branch: str, base: str, new: str) -> str:
|
||||
self.created.append((branch, base, new))
|
||||
return new
|
||||
|
||||
def compare_and_swap_branch(self, *_args: object) -> str:
|
||||
raise AssertionError("existing-branch CAS must not serve CREATE_NEW")
|
||||
|
||||
class CreateRemote:
|
||||
def __init__(self) -> None:
|
||||
self.created: list[tuple[str, str]] = []
|
||||
|
||||
def push_new_exact(self, branch: str, new: str) -> None:
|
||||
self.created.append((branch, new))
|
||||
|
||||
def push_exact(self, *_args: object) -> None:
|
||||
raise AssertionError("existing-branch push must not serve CREATE_NEW")
|
||||
|
||||
class Credential:
|
||||
def project(self) -> None: pass
|
||||
def revoke(self) -> None: pass
|
||||
|
||||
controller=Controller(); authority=CreateAuthority(); remote=CreateRemote()
|
||||
executor=ContinuationExecutor(controller,authority,Workspace(),remote,Credential())
|
||||
executor._validate_preflight=lambda *_args,**_kwargs: None
|
||||
executor._finish_remote_verification=lambda *_args,**_kwargs: {
|
||||
"job_id":JOB,"state":"COMPLETED","receipt":{"authority_oid":new_oid,"commit_visible":True}
|
||||
}
|
||||
now=datetime.now(timezone.utc)
|
||||
binding=ContinuationRuntimeBinding(
|
||||
request=lease(branch_mode="CREATE_NEW"),job_id=JOB,
|
||||
workspace_path=Path("/tmp/create-new-executor-resume"),
|
||||
lease_not_before=now-timedelta(seconds=1),lease_expires_at=now+timedelta(seconds=60),
|
||||
original_receipt_valid=True,original_grant_consumed=True,
|
||||
permissions=ContinuationPermissions(),
|
||||
)
|
||||
result=executor.execute(
|
||||
binding,ContinuationCommit("create-new-test",{"docs/test.txt":b"ok"},"test")
|
||||
)
|
||||
|
||||
assert result["state"]=="COMPLETED"
|
||||
assert authority.created==[(BRANCH,BASE,new_oid)]
|
||||
assert remote.created==[(BRANCH,new_oid)]
|
||||
assert controller.advanced==["LOCAL_VERIFIED","REMOTE_SYNCING"]
|
||||
@@ -1,763 +0,0 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from concurrent.futures import ThreadPoolExecutor
|
||||
from datetime import datetime, timedelta, timezone
|
||||
import hashlib
|
||||
from dataclasses import asdict
|
||||
from pathlib import Path
|
||||
import sqlite3
|
||||
import json
|
||||
from types import SimpleNamespace
|
||||
import threading
|
||||
|
||||
import pytest
|
||||
|
||||
from aios_chatgpt_bridge_v2_server import GovernedRuntime
|
||||
from aios_continuation_executor import ContinuationPermissions, ContinuationRuntimeBinding
|
||||
from aios_continuation_lease import WriterLeaseRequestV2
|
||||
from aios_gitops_job_controller import GitOpsJobController, JobError, JobStore
|
||||
from aios_governed_task_bootstrap import GovernedTaskBootstrap
|
||||
|
||||
|
||||
TASK_ID = "gov-writer-lease-recovery-test"
|
||||
BRANCH = "agent/writer-lease-recovery-test"
|
||||
BASE = "a" * 40
|
||||
JOB_ID_PREFIX = "writer-lease-runtime-lineage"
|
||||
|
||||
|
||||
class LocalAuthority:
|
||||
def __init__(self, head: str | None = None) -> None:
|
||||
self.head = head
|
||||
|
||||
def read_ref(self, _ref: str) -> str | None:
|
||||
return self.head
|
||||
|
||||
def commit_exists(self, _oid: str) -> bool:
|
||||
return True
|
||||
|
||||
def ancestry_check(self, _older: str, _newer: str) -> bool:
|
||||
return True
|
||||
|
||||
|
||||
def writer_request(task_id: str = TASK_ID, **overrides: object) -> dict[str, object]:
|
||||
request: dict[str, object] = {
|
||||
"schema": WriterLeaseRequestV2.schema,
|
||||
"task_id": task_id,
|
||||
"repository": "aios/aios",
|
||||
"branch_mode": "CREATE_NEW",
|
||||
"base_oid": BASE,
|
||||
"expected_branch_oid": BASE,
|
||||
"target_branch": BRANCH,
|
||||
"generation": 1,
|
||||
"lease_seconds": 120,
|
||||
}
|
||||
request.update(overrides)
|
||||
return request
|
||||
|
||||
|
||||
def make_runtime(tmp_path: Path, task_id: str = TASK_ID, *, authority_head: str | None = None) -> GovernedRuntime:
|
||||
runtime = GovernedRuntime.__new__(GovernedRuntime)
|
||||
runtime.config = SimpleNamespace(repository="aios/aios", state_root=tmp_path)
|
||||
runtime.authority = LocalAuthority(authority_head)
|
||||
runtime.bindings = {
|
||||
task_id: {
|
||||
"task_id": task_id,
|
||||
"repository": "aios/aios",
|
||||
"base_oid": BASE,
|
||||
"expected_branch_oid": BASE,
|
||||
"target_branch": BRANCH,
|
||||
"branch_mode": "CREATE_NEW",
|
||||
"generation": 1,
|
||||
}
|
||||
}
|
||||
runtime.runtimes = {}
|
||||
runtime.controllers = {}
|
||||
runtime.bootstrap = GovernedTaskBootstrap(
|
||||
runtime.authority,
|
||||
tmp_path,
|
||||
remote_name="test",
|
||||
token_path=tmp_path / "token",
|
||||
ssh_key_path=tmp_path / "key",
|
||||
)
|
||||
runtime.bindings = {}
|
||||
runtime.create_task(
|
||||
{
|
||||
"task_id": task_id,
|
||||
"repository": "aios/aios",
|
||||
"base_oid": BASE,
|
||||
"expected_branch_oid": BASE,
|
||||
"target_branch": BRANCH,
|
||||
"risk": "engineering",
|
||||
"description": "governed runtime lineage test",
|
||||
"idempotency_key": JOB_ID_PREFIX + "-" + task_id,
|
||||
}
|
||||
)
|
||||
job_root = runtime._job_root(task_id)
|
||||
runtime.controllers[task_id] = GitOpsJobController(
|
||||
JobStore(job_root, payload_root=job_root / "payloads")
|
||||
)
|
||||
|
||||
def build(
|
||||
lease: WriterLeaseRequestV2,
|
||||
*,
|
||||
job_id: str,
|
||||
lease_not_before: datetime | None = None,
|
||||
lease_expires_at: datetime | None = None,
|
||||
) -> tuple[ContinuationRuntimeBinding, GitOpsJobController, object]:
|
||||
controller = runtime.controllers.get(lease.task_id)
|
||||
if controller is None:
|
||||
root = runtime._job_root(lease.task_id)
|
||||
controller = GitOpsJobController(JobStore(root, payload_root=root / "payloads"))
|
||||
now = datetime.now(timezone.utc)
|
||||
return (
|
||||
ContinuationRuntimeBinding(
|
||||
request=lease,
|
||||
job_id=job_id,
|
||||
workspace_path=tmp_path / "worktrees" / lease.task_id,
|
||||
lease_not_before=lease_not_before or now - timedelta(seconds=1),
|
||||
lease_expires_at=lease_expires_at or now + timedelta(seconds=lease.lease_seconds),
|
||||
original_receipt_valid=True,
|
||||
original_grant_consumed=True,
|
||||
permissions=ContinuationPermissions(),
|
||||
),
|
||||
controller,
|
||||
object(),
|
||||
)
|
||||
|
||||
runtime._build_runtime = build
|
||||
return runtime
|
||||
|
||||
|
||||
def create_durable_lease(
|
||||
runtime: GovernedRuntime,
|
||||
*,
|
||||
expires_at: str = "2020-01-01T00:00:00+00:00",
|
||||
) -> GitOpsJobController:
|
||||
root = runtime._job_root(TASK_ID)
|
||||
controller = GitOpsJobController(JobStore(root, payload_root=root / "payloads"))
|
||||
receipt = SimpleNamespace(job_id=runtime.bindings[TASK_ID]["job_id"])
|
||||
controller.lease_job(
|
||||
receipt.job_id,
|
||||
lease_id="writer-lease-recovery",
|
||||
lineage_anchor_oid=BASE,
|
||||
current_head_oid=BASE,
|
||||
generation=1,
|
||||
lease_expires_at=expires_at,
|
||||
expected_revision=0,
|
||||
)
|
||||
runtime.controllers[TASK_ID] = controller
|
||||
return controller
|
||||
|
||||
|
||||
def durable_row(runtime: GovernedRuntime) -> tuple[str, int]:
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
row = connection.execute(
|
||||
"SELECT lease_expires_at, revision FROM branch_leases WHERE lease_id=?",
|
||||
("writer-lease-recovery",),
|
||||
).fetchone()
|
||||
assert row is not None
|
||||
return row
|
||||
|
||||
|
||||
def test_first_bootstrap_refreshes_only_runtime_without_durable_lease(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
result = runtime.bootstrap_writer_lease(writer_request())
|
||||
|
||||
assert result["bootstrapped"] is True
|
||||
assert result["runtime_rehydrated"] is False
|
||||
assert result["durable_branch_lease_present"] is False
|
||||
assert result["idempotent"] is False
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
assert connection.execute("SELECT COUNT(*) FROM branch_leases").fetchone()[0] == 0
|
||||
|
||||
|
||||
def test_duplicate_bootstrap_is_idempotent_without_consuming_generation(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
first = runtime.bootstrap_writer_lease(writer_request())
|
||||
second = runtime.bootstrap_writer_lease(writer_request())
|
||||
|
||||
assert first["bootstrapped"] is True
|
||||
assert first["generation"] == 1
|
||||
assert second["bootstrapped"] is False
|
||||
assert second["idempotent"] is True
|
||||
assert second["generation"] == 1
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
assert connection.execute("SELECT COUNT(*) FROM branch_leases").fetchone()[0] == 0
|
||||
|
||||
|
||||
def test_commit_after_bootstrap_reuses_the_exact_runtime_binding(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
class RecordingWorkspace:
|
||||
|
||||
def __init__(self) -> None:
|
||||
self.files: dict[str, bytes] = {}
|
||||
|
||||
def materialize_files(self, files: dict[str, bytes]) -> None:
|
||||
self.files = dict(files)
|
||||
|
||||
class RecordingExecutor:
|
||||
|
||||
def __init__(self) -> None:
|
||||
self.workspace = RecordingWorkspace()
|
||||
self.binding = None
|
||||
self.commit = None
|
||||
|
||||
def execute(self, binding: object, commit: object) -> dict[str, object]:
|
||||
self.binding = binding
|
||||
self.commit = commit
|
||||
return {'task_id': binding.request.task_id, 'generation': binding.request.generation, 'files': sorted(self.workspace.files)}
|
||||
executor = RecordingExecutor()
|
||||
original_build = runtime._build_runtime
|
||||
|
||||
def build(lease: WriterLeaseRequestV2, *, job_id: str, lease_not_before: datetime | None=None, lease_expires_at: datetime | None=None) -> tuple[ContinuationRuntimeBinding, GitOpsJobController, object]:
|
||||
binding, controller, _ = original_build(lease, job_id=job_id, lease_not_before=lease_not_before, lease_expires_at=lease_expires_at)
|
||||
return (binding, controller, executor)
|
||||
runtime._build_runtime = build
|
||||
runtime.bootstrap_writer_lease(writer_request())
|
||||
lease_response = runtime.acquire_lease(writer_request())
|
||||
binding, controller, executor = runtime.runtimes[TASK_ID, 1]
|
||||
assert lease_response['job_id'] == binding.job_id
|
||||
assert lease_response['durable_branch_lease_present'] is True
|
||||
_original_execute = executor.execute
|
||||
|
||||
def _execute_with_governed_commit_contract(*args, **kwargs):
|
||||
_before_complete = kwargs.pop('before_complete', None)
|
||||
_legacy = dict(_original_execute(*args, **kwargs))
|
||||
_result = {**_legacy, 'job_id': binding.job_id, 'state': 'COMPLETED', 'side_effect_state': 'REMOTE_VERIFIED', 'receipt': {'authority_oid': 'b' * 40, 'git_remote_oid': 'b' * 40, 'api_branch_oid': 'b' * 40, 'commit_visible': True, 'remote_verified': True, 'force_push': False}}
|
||||
if _before_complete is not None:
|
||||
_before_complete('b' * 40)
|
||||
return _result
|
||||
executor.execute = _execute_with_governed_commit_contract
|
||||
_bound_job_id = binding.job_id
|
||||
result = runtime.commit_files({'task_id': TASK_ID, 'repository': 'aios/aios', 'generation': 1, 'target_branch': BRANCH, 'expected_branch_oid': BASE, 'idempotency_key': 'commit-after-bootstrap-v1', 'path': 'docs/recovery.txt', 'content_base64': 'YQ==', 'message': 'test commit after bootstrap'})
|
||||
assert result['task_id'] == TASK_ID
|
||||
assert result['generation'] == 1
|
||||
assert result['files'] == ['docs/recovery.txt']
|
||||
assert executor.binding.request.task_id == TASK_ID
|
||||
assert executor.binding.request.generation == 1
|
||||
database = runtime._job_root(TASK_ID) / 'state.sqlite3'
|
||||
with sqlite3.connect(f'file:{database}?mode=ro', uri=True) as connection:
|
||||
assert connection.execute('SELECT COUNT(*) FROM branch_leases').fetchone()[0] == 1
|
||||
assert connection.execute('PRAGMA foreign_key_check(branch_leases)').fetchall() == []
|
||||
assert result['commit_oid'] == 'b' * 40
|
||||
assert result['commit_visible'] is True
|
||||
assert result['branch_mode'] == 'CONTINUE_EXISTING'
|
||||
assert result['target_branch'] == BRANCH
|
||||
assert int(result['generation']) == 1
|
||||
with controller.store._connect() as connection:
|
||||
_persisted = connection.execute('SELECT current_head_oid,revision FROM branch_leases WHERE lease_id=?', ('lease-' + str(_bound_job_id),)).fetchone()
|
||||
assert _persisted is not None
|
||||
assert _persisted[0] == 'b' * 40
|
||||
assert int(_persisted[1]) == 1
|
||||
fresh_runtime = make_runtime(tmp_path, authority_head='b' * 40)
|
||||
fresh_response = fresh_runtime.bootstrap_writer_lease(writer_request(branch_mode='CONTINUE_EXISTING', expected_branch_oid='b' * 40))
|
||||
fresh_binding, _, _ = fresh_runtime.runtimes[TASK_ID, 1]
|
||||
assert fresh_response['job_id'] == str(_bound_job_id)
|
||||
assert fresh_response['runtime_rehydrated'] is True
|
||||
assert fresh_binding.job_id == str(_bound_job_id)
|
||||
assert fresh_binding.request.expected_branch_oid == 'b' * 40
|
||||
with fresh_runtime.controllers[TASK_ID].store._connect() as connection:
|
||||
_fresh_lease = connection.execute('SELECT job_id,current_head_oid,revision FROM branch_leases WHERE lease_id=?', ('lease-' + str(_bound_job_id),)).fetchone()
|
||||
assert tuple(_fresh_lease) == (str(_bound_job_id), 'b' * 40, 1)
|
||||
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
("field", "value", "error"),
|
||||
(
|
||||
("repository", "other/repository", "LEASE_BINDING_CONFLICT"),
|
||||
("base_oid", "b" * 40, "LEASE_BINDING_CONFLICT"),
|
||||
("expected_branch_oid", "b" * 40, "LEASE_BINDING_CONFLICT"),
|
||||
("target_branch", "agent/wrong-target", "LEASE_BINDING_CONFLICT"),
|
||||
("generation", 2, "GENERATION_BINDING_CONFLICT"),
|
||||
),
|
||||
)
|
||||
def test_bootstrap_rejects_any_nonmatching_binding_field(
|
||||
tmp_path: Path, field: str, value: object, error: str
|
||||
) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
with pytest.raises(RuntimeError, match=error):
|
||||
runtime.bootstrap_writer_lease(writer_request(**{field: value}))
|
||||
|
||||
|
||||
def test_bootstrap_rejects_unknown_task_ownership(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
with pytest.raises(RuntimeError, match="TASK_OWNERSHIP_INVALID"):
|
||||
runtime.bootstrap_writer_lease(writer_request(task_id="unknown-task"))
|
||||
|
||||
|
||||
def test_renew_missing_durable_lease_is_not_bootstrap(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
with pytest.raises(RuntimeError, match="LEASE_NOT_FOUND"):
|
||||
runtime.renew_writer_lease(writer_request())
|
||||
|
||||
|
||||
def test_renew_is_generation_scoped_before_durable_lookup(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
create_durable_lease(runtime)
|
||||
|
||||
with pytest.raises(RuntimeError, match="GENERATION_BINDING_CONFLICT"):
|
||||
runtime.renew_writer_lease(writer_request(generation=2))
|
||||
assert durable_row(runtime) == ("2020-01-01T00:00:00+00:00", 0)
|
||||
|
||||
|
||||
def test_restart_rehydrates_only_from_matching_durable_lease(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
create_durable_lease(runtime)
|
||||
runtime.runtimes.clear()
|
||||
|
||||
result = runtime.renew_writer_lease(writer_request())
|
||||
|
||||
expires_at, revision = durable_row(runtime)
|
||||
assert result["recovered"] is True
|
||||
assert result["runtime_rehydrated"] is True
|
||||
assert result["durable_branch_lease_present"] is True
|
||||
assert revision == 1
|
||||
assert datetime.fromisoformat(expires_at) > datetime.now(timezone.utc)
|
||||
|
||||
|
||||
def test_duplicate_renewal_is_idempotent_after_first_durable_recovery(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
create_durable_lease(runtime)
|
||||
|
||||
first = runtime.renew_writer_lease(writer_request())
|
||||
second = runtime.renew_writer_lease(writer_request())
|
||||
|
||||
assert first["recovered"] is True
|
||||
assert first["idempotent"] is False
|
||||
assert second["recovered"] is False
|
||||
assert second["idempotent"] is True
|
||||
assert durable_row(runtime)[1] == 1
|
||||
|
||||
|
||||
def test_update_lease_keeps_expiry_without_optional_argument_and_keeps_cas(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
controller = create_durable_lease(runtime)
|
||||
|
||||
controller.update_lease(
|
||||
"writer-lease-recovery",
|
||||
expected_revision=0,
|
||||
current_head_oid=BASE,
|
||||
generation=1,
|
||||
)
|
||||
assert durable_row(runtime) == ("2020-01-01T00:00:00+00:00", 1)
|
||||
with pytest.raises(JobError, match="LEASE_REVISION_CAS_CONFLICT"):
|
||||
controller.update_lease(
|
||||
"writer-lease-recovery",
|
||||
expected_revision=0,
|
||||
current_head_oid=BASE,
|
||||
generation=1,
|
||||
)
|
||||
|
||||
|
||||
def test_two_runtime_connections_resolve_renewal_cas_race_idempotently(tmp_path: Path) -> None:
|
||||
first = make_runtime(tmp_path)
|
||||
first_controller = create_durable_lease(first)
|
||||
second = make_runtime(tmp_path)
|
||||
second_controller = GitOpsJobController(
|
||||
JobStore(second._job_root(TASK_ID), payload_root=second._job_root(TASK_ID) / "payloads")
|
||||
)
|
||||
second.controllers[TASK_ID] = second_controller
|
||||
gate = threading.Barrier(2)
|
||||
|
||||
def gate_update(controller: GitOpsJobController) -> None:
|
||||
original = controller.update_lease
|
||||
|
||||
def update(*args: object, **kwargs: object) -> None:
|
||||
gate.wait(timeout=5)
|
||||
original(*args, **kwargs)
|
||||
|
||||
controller.update_lease = update
|
||||
|
||||
gate_update(first_controller)
|
||||
gate_update(second_controller)
|
||||
with ThreadPoolExecutor(max_workers=2) as pool:
|
||||
results = list(pool.map(lambda runtime: runtime.renew_writer_lease(writer_request()), (first, second)))
|
||||
|
||||
assert sorted((item["recovered"], item["idempotent"]) for item in results) == [
|
||||
(False, True),
|
||||
(True, False),
|
||||
]
|
||||
assert durable_row(first)[1] == 1
|
||||
|
||||
|
||||
def test_bootstrap_preserves_exact_governed_job_lineage(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
|
||||
response = runtime.bootstrap_writer_lease(writer_request())
|
||||
binding, _, _ = runtime.runtimes[(TASK_ID, 1)]
|
||||
|
||||
assert response["job_id"] == expected_job_id
|
||||
assert binding.job_id == expected_job_id
|
||||
assert binding.request.job_id == expected_job_id
|
||||
|
||||
|
||||
def test_restart_rehydrate_preserves_exact_governed_job_lineage(tmp_path: Path) -> None:
|
||||
first = make_runtime(tmp_path)
|
||||
expected_job_id = first.bindings[TASK_ID]["job_id"]
|
||||
first.acquire_lease(writer_request())
|
||||
|
||||
runtime = make_runtime(tmp_path)
|
||||
assert runtime is not first
|
||||
assert runtime.runtimes == {}
|
||||
response = runtime.bootstrap_writer_lease(writer_request())
|
||||
binding, _, _ = runtime.runtimes[(TASK_ID, 1)]
|
||||
|
||||
assert response["job_id"] == expected_job_id
|
||||
assert binding.job_id == expected_job_id
|
||||
assert binding.request.job_id == expected_job_id
|
||||
|
||||
|
||||
def test_wrong_governed_job_id_is_denied_before_runtime_or_lease_side_effect(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
|
||||
with pytest.raises(RuntimeError, match="LEASE_BINDING_CONFLICT"):
|
||||
runtime.bootstrap_writer_lease(writer_request(job_id="job-wrong-lineage"))
|
||||
|
||||
assert runtime.runtimes == {}
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
assert connection.execute("SELECT COUNT(*) FROM branch_leases").fetchone()[0] == 0
|
||||
assert connection.execute("PRAGMA foreign_key_check(branch_leases)").fetchall() == []
|
||||
|
||||
|
||||
def test_acquire_creates_exact_governed_writer_lease_with_fk_and_cas(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
|
||||
response = runtime.acquire_lease(writer_request())
|
||||
|
||||
assert response["job_id"] == expected_job_id
|
||||
assert response["durable_branch_lease_present"] is True
|
||||
assert response["idempotent"] is False
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
foreign_keys = connection.execute(
|
||||
"PRAGMA foreign_key_list(branch_leases)"
|
||||
).fetchall()
|
||||
row = connection.execute(
|
||||
"SELECT job_id,branch,lineage_anchor_oid,current_head_oid,generation,state,revision "
|
||||
"FROM branch_leases"
|
||||
).fetchone()
|
||||
job = connection.execute(
|
||||
"SELECT state,side_effect_state,revision FROM jobs WHERE job_id=?",
|
||||
(expected_job_id,),
|
||||
).fetchone()
|
||||
assert connection.execute("PRAGMA foreign_key_check(branch_leases)").fetchall() == []
|
||||
assert any(
|
||||
fk[2] == "jobs" and fk[3] == "job_id" and fk[4] == "job_id"
|
||||
for fk in foreign_keys
|
||||
)
|
||||
assert row == (expected_job_id, BRANCH, BASE, BASE, 1, "ACTIVE", 0)
|
||||
assert job == ("LEASED", "LEASED", 2)
|
||||
|
||||
|
||||
def test_duplicate_acquire_observes_same_durable_writer_lease(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
|
||||
first = runtime.acquire_lease(writer_request())
|
||||
second = runtime.acquire_lease(writer_request())
|
||||
|
||||
assert first["job_id"] == expected_job_id
|
||||
assert second["job_id"] == expected_job_id
|
||||
assert second["idempotent"] is True
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
leases = connection.execute(
|
||||
"SELECT lease_id,job_id,revision FROM branch_leases"
|
||||
).fetchall()
|
||||
job = connection.execute(
|
||||
"SELECT state,revision FROM jobs WHERE job_id=?",
|
||||
(expected_job_id,),
|
||||
).fetchone()
|
||||
assert leases == [("lease-" + expected_job_id, expected_job_id, 0)]
|
||||
assert job == ("LEASED", 2)
|
||||
|
||||
|
||||
def test_wrong_job_acquire_is_denied_before_authorize_or_lease(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
|
||||
with pytest.raises(RuntimeError, match="LEASE_BINDING_CONFLICT"):
|
||||
runtime.acquire_lease(writer_request(job_id="job-wrong-lineage"))
|
||||
|
||||
database = runtime._job_root(TASK_ID) / "state.sqlite3"
|
||||
with sqlite3.connect(f"file:{database}?mode=ro", uri=True) as connection:
|
||||
job = connection.execute(
|
||||
"SELECT state,side_effect_state,revision FROM jobs WHERE job_id=?",
|
||||
(expected_job_id,),
|
||||
).fetchone()
|
||||
lease_count = connection.execute("SELECT COUNT(*) FROM branch_leases").fetchone()[0]
|
||||
assert job == ("REQUESTED", "NONE", 0)
|
||||
assert lease_count == 0
|
||||
|
||||
|
||||
def test_omitted_request_job_id_resolves_the_exact_durable_parent(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
raw = dict(writer_request())
|
||||
|
||||
response = runtime.acquire_lease(raw)
|
||||
binding, _, _ = runtime.runtimes[(TASK_ID, 1)]
|
||||
|
||||
assert response["job_id"] == expected_job_id
|
||||
assert binding.job_id == expected_job_id
|
||||
assert binding.request.job_id == expected_job_id
|
||||
|
||||
|
||||
def test_durable_job_expected_head_mismatch_is_denied(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
expected_job_id = runtime.bindings[TASK_ID]["job_id"]
|
||||
controller = runtime.controllers[TASK_ID]
|
||||
with controller.store._connect() as connection:
|
||||
row = connection.execute(
|
||||
"SELECT request_json FROM jobs WHERE job_id=?", (expected_job_id,)
|
||||
).fetchone()
|
||||
request = json.loads(row[0])
|
||||
request["expected_branch_oid"] = "c" * 40
|
||||
request["binding"]["expected_branch_oid"] = "c" * 40
|
||||
connection.execute(
|
||||
"UPDATE jobs SET request_json=? WHERE job_id=?",
|
||||
(json.dumps(request, sort_keys=True, separators=(",", ":")), expected_job_id),
|
||||
)
|
||||
connection.commit()
|
||||
|
||||
with pytest.raises(RuntimeError, match="RUNTIME_GOVERNED_JOB_UNRESOLVED"):
|
||||
runtime.bootstrap_writer_lease(writer_request())
|
||||
|
||||
assert runtime.runtimes == {}
|
||||
|
||||
|
||||
def _commit_ready_runtime(tmp_path: Path):
|
||||
runtime = make_runtime(tmp_path)
|
||||
runtime.bootstrap_writer_lease(writer_request())
|
||||
binding, controller, _ = runtime.runtimes[(TASK_ID, 1)]
|
||||
|
||||
class Workspace:
|
||||
def materialize_files(self, _files):
|
||||
return None
|
||||
|
||||
class Executor:
|
||||
def __init__(self):
|
||||
self.workspace = Workspace()
|
||||
self.calls = 0
|
||||
|
||||
def execute(self, _binding, _commit, *, before_complete=None):
|
||||
self.calls += 1
|
||||
result = {
|
||||
"job_id": binding.job_id,
|
||||
"state": "COMPLETED",
|
||||
"side_effect_state": "REMOTE_VERIFIED",
|
||||
"receipt": {
|
||||
"authority_oid": "b" * 40,
|
||||
"git_remote_oid": "b" * 40,
|
||||
"api_branch_oid": "b" * 40,
|
||||
"commit_visible": True,
|
||||
"remote_verified": True,
|
||||
"force_push": False,
|
||||
},
|
||||
}
|
||||
if before_complete is not None:
|
||||
before_complete("b" * 40)
|
||||
return result
|
||||
|
||||
executor = Executor()
|
||||
runtime.runtimes[(TASK_ID, 1)] = (binding, controller, executor)
|
||||
runtime.acquire_lease(writer_request())
|
||||
return runtime, binding, controller, executor
|
||||
|
||||
|
||||
def _commit_request():
|
||||
return {
|
||||
"task_id": TASK_ID,
|
||||
"repository": "aios/aios",
|
||||
"generation": 1,
|
||||
"target_branch": BRANCH,
|
||||
"expected_branch_oid": BASE,
|
||||
"idempotency_key": "commit-cas-test-v1",
|
||||
"path": "docs/cas.txt",
|
||||
"content_base64": "YQ==",
|
||||
"message": "test commit cas",
|
||||
}
|
||||
|
||||
|
||||
def test_commit_lease_revision_cas_conflict_is_preserved(tmp_path: Path) -> None:
|
||||
runtime, _binding, controller, _executor = _commit_ready_runtime(tmp_path)
|
||||
|
||||
def conflict(*_args, **_kwargs):
|
||||
raise JobError("LEASE_REVISION_CAS_CONFLICT")
|
||||
|
||||
controller.update_lease = conflict
|
||||
with pytest.raises(RuntimeError, match="LEASE_CAS_CONFLICT"):
|
||||
runtime.commit_files(_commit_request())
|
||||
|
||||
|
||||
def test_commit_divergent_durable_head_fails_before_executor(tmp_path: Path) -> None:
|
||||
runtime, binding, controller, executor = _commit_ready_runtime(tmp_path)
|
||||
controller.update_lease(
|
||||
"lease-" + binding.job_id,
|
||||
expected_revision=0,
|
||||
current_head_oid="c" * 40,
|
||||
generation=1,
|
||||
)
|
||||
|
||||
with pytest.raises(RuntimeError, match="LEASE_HEAD_RECONCILIATION_REQUIRED"):
|
||||
runtime.commit_files(_commit_request())
|
||||
assert executor.calls == 0
|
||||
|
||||
|
||||
def test_remote_pending_never_advances_durable_head(tmp_path: Path) -> None:
|
||||
runtime, binding, controller, executor = _commit_ready_runtime(tmp_path)
|
||||
|
||||
def pending(_binding, _commit, **_kwargs):
|
||||
executor.calls += 1
|
||||
return {
|
||||
"job_id": binding.job_id,
|
||||
"state": "RECONCILIATION_REQUIRED",
|
||||
"side_effect_state": "REMOTE_PUSH_PENDING",
|
||||
"receipt": {
|
||||
"authority_oid": "b" * 40,
|
||||
"desired_remote_oid": "b" * 40,
|
||||
"remote_verified": False,
|
||||
},
|
||||
}
|
||||
|
||||
executor.execute = pending
|
||||
with pytest.raises(RuntimeError, match="RECONCILIATION_REQUIRED"):
|
||||
runtime.commit_files(_commit_request())
|
||||
with controller.store._connect() as connection:
|
||||
row = connection.execute(
|
||||
"SELECT current_head_oid,revision FROM branch_leases WHERE lease_id=?",
|
||||
("lease-" + binding.job_id,),
|
||||
).fetchone()
|
||||
assert tuple(row) == (BASE, 0)
|
||||
|
||||
|
||||
def test_retry_accepts_exact_head_already_advanced_at_remote_verified_boundary(tmp_path: Path) -> None:
|
||||
runtime, binding, controller, executor = _commit_ready_runtime(tmp_path)
|
||||
commit_oid = "b" * 40
|
||||
controller.update_lease(
|
||||
"lease-" + binding.job_id,
|
||||
expected_revision=0,
|
||||
current_head_oid=commit_oid,
|
||||
generation=1,
|
||||
)
|
||||
receipt = {
|
||||
"authority_oid": commit_oid,
|
||||
"git_remote_oid": commit_oid,
|
||||
"api_branch_oid": commit_oid,
|
||||
"commit_visible": True,
|
||||
"remote_verified": True,
|
||||
"force_push": False,
|
||||
}
|
||||
with controller.store._connect() as connection:
|
||||
connection.execute(
|
||||
"UPDATE jobs SET state='REMOTE_VERIFIED',side_effect_state='REMOTE_VERIFIED',"
|
||||
"receipt_json=? WHERE job_id=?",
|
||||
(json.dumps(receipt, sort_keys=True, separators=(",", ":")), binding.job_id),
|
||||
)
|
||||
connection.commit()
|
||||
|
||||
result = runtime.commit_files(_commit_request())
|
||||
|
||||
assert result["commit_oid"] == commit_oid
|
||||
assert executor.calls == 1
|
||||
with controller.store._connect() as connection:
|
||||
row = connection.execute(
|
||||
"SELECT current_head_oid,revision FROM branch_leases WHERE lease_id=?",
|
||||
("lease-" + binding.job_id,),
|
||||
).fetchone()
|
||||
assert tuple(row) == (commit_oid, 1)
|
||||
|
||||
|
||||
def test_completed_response_with_invalid_oid_uses_canonical_oid_error(tmp_path: Path) -> None:
|
||||
runtime, binding, controller, executor = _commit_ready_runtime(tmp_path)
|
||||
|
||||
def invalid_oid(_binding, _commit, **_kwargs):
|
||||
executor.calls += 1
|
||||
return {
|
||||
"job_id": binding.job_id,
|
||||
"state": "COMPLETED",
|
||||
"side_effect_state": "REMOTE_VERIFIED",
|
||||
"receipt": {
|
||||
"authority_oid": "not-an-oid",
|
||||
"commit_visible": True,
|
||||
"remote_verified": True,
|
||||
},
|
||||
}
|
||||
|
||||
executor.execute = invalid_oid
|
||||
with pytest.raises(RuntimeError, match="GOVERNED_COMMIT_OID_INVALID"):
|
||||
runtime.commit_files(_commit_request())
|
||||
with controller.store._connect() as connection:
|
||||
row = connection.execute(
|
||||
"SELECT current_head_oid,revision FROM branch_leases WHERE lease_id=?",
|
||||
("lease-" + binding.job_id,),
|
||||
).fetchone()
|
||||
assert tuple(row) == (BASE, 0)
|
||||
|
||||
|
||||
def test_acquire_renews_an_expired_matching_durable_lease_with_revision_cas(tmp_path: Path) -> None:
|
||||
runtime = make_runtime(tmp_path)
|
||||
create_durable_lease(runtime)
|
||||
runtime.runtimes.clear()
|
||||
|
||||
result = runtime.acquire_lease(writer_request())
|
||||
|
||||
expires_at, revision = durable_row(runtime)
|
||||
assert result["recovered"] is True
|
||||
assert result["runtime_rehydrated"] is False
|
||||
assert result["durable_branch_lease_present"] is True
|
||||
assert revision == 1
|
||||
assert datetime.fromisoformat(expires_at) > datetime.now(timezone.utc)
|
||||
|
||||
|
||||
def test_authority_recovers_only_clean_detached_single_attempt_worktree(tmp_path: Path) -> None:
|
||||
import subprocess
|
||||
from aios_authority_git_executor import AuthorityGitExecutor
|
||||
|
||||
bare = tmp_path / "authority.git"
|
||||
seed = tmp_path / "seed"
|
||||
subprocess.run(["git", "init", "--bare", str(bare)], check=True, capture_output=True)
|
||||
subprocess.run(["git", "init", str(seed)], check=True, capture_output=True)
|
||||
(seed / "base.txt").write_text("base\n", encoding="utf-8")
|
||||
subprocess.run(["git", "-C", str(seed), "add", "base.txt"], check=True)
|
||||
subprocess.run(
|
||||
["git", "-C", str(seed), "-c", "user.name=test", "-c", "user.email=test@example.invalid",
|
||||
"commit", "-m", "base"], check=True, capture_output=True,
|
||||
)
|
||||
base = subprocess.check_output(["git", "-C", str(seed), "rev-parse", "HEAD"], text=True).strip()
|
||||
subprocess.run(["git", "-C", str(seed), "push", str(bare), "HEAD:refs/heads/main"], check=True, capture_output=True)
|
||||
root = tmp_path / "worktrees"
|
||||
authority = AuthorityGitExecutor(bare, root)
|
||||
worktree = authority.create_detached_worktree(base, "task")
|
||||
(worktree / "attempt.txt").write_text("attempt\n", encoding="utf-8")
|
||||
subprocess.run(["git", "-C", str(worktree), "add", "attempt.txt"], check=True)
|
||||
subprocess.run(
|
||||
["git", "-C", str(worktree), "-c", "user.name=test", "-c", "user.email=test@example.invalid",
|
||||
"commit", "-m", "failed attempt"], check=True, capture_output=True,
|
||||
)
|
||||
stale = subprocess.check_output(["git", "-C", str(worktree), "rev-parse", "HEAD"], text=True).strip()
|
||||
|
||||
recovered = authority.recover_failed_detached_worktree_exact(
|
||||
path=worktree,
|
||||
expected_oid=base,
|
||||
stale_oid=stale,
|
||||
target_ref="refs/heads/agent/recovery-test",
|
||||
)
|
||||
|
||||
assert recovered == stale
|
||||
assert subprocess.check_output(["git", "-C", str(worktree), "rev-parse", "HEAD"], text=True).strip() == base
|
||||
assert subprocess.run(["git", "-C", str(worktree), "symbolic-ref", "-q", "HEAD"]).returncode == 1
|
||||
assert subprocess.check_output(["git", "-C", str(worktree), "status", "--porcelain"], text=True) == ""
|
||||
Reference in New Issue
Block a user