Manager

Autonomous agent manager loop.

Polls active user tasks, runs a single pipeline iteration per task, persists results (unless dry-run), and triggers Telegram notifications via DB tasks.

class agent.manager.RuntimeConfig(poll_seconds=30, dry_run=False, agent_id='main_agent', test_user_id=None)

Bases: object

In-memory configuration for the agent manager.

Variables:

• poll_seconds – Poll interval in seconds for the main loop.

• dry_run – If True, do not persist analyses; only notify.

• agent_id – Identifier reported in status updates.

• test_user_id – Optional override to send notifications to a test user.

Parameters:

• poll_seconds (int)

• dry_run (bool)

• agent_id (str)

• test_user_id (int | None)

agent_id: str = 'main_agent'

dry_run: bool = False

poll_seconds: int = 30

test_user_id: Optional[int] = None

async agent.manager.main()

Agent main loop: poll tasks and process them autonomously.

Return type:

None

Returns:

None.