You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
AGENTS.md still says the controller runs every 30 minutes #1197
After #1194, the label controller is event-driven: it runs on pull_requestopened, ready_for_review and reopened (same-repo; a fork-head run is skipped because that token cannot write) and on a four-a-day cron backstop (17 */6 * * *).
Two sentences in AGENTS.md still describe the old half-hourly schedule:
Merge policy: "the controller sets the risk:T* label within ~30 minutes."
Controller section: it "runs every 30 minutes".
The module docstring in tests/test_process_controller.py repeats "It runs every 30 minutes…".
Agents reading the process doc will wait for a 30-minute tick that no longer exists.
Expected
Those three spots describe the live triggers: same-repo PRs are labelled on the lifecycle events; fork PRs (and any missed event) wait for the four-a-day cron backstop. No change to controller.yml or controller.py.
Acceptance
AGENTS.md no longer says the controller runs every 30 minutes or sets risk:T* "within ~30 minutes".
Activity
Birfy commentedon Nov 20, 2026
Claiming as Birfy on branch
agent/1197-controller-schedule-docs.