io.github.dingdawg/dingdawg-loop
Safe scheduled AI agents with governance gates. Verified, receipted, fail-closed.
Versions
2.0.8latestTools 6
register_loop Register a scheduled agent loop. Stores the loop definition with its cron schedule, action type, and risk tier. The loop is created in
execute_loop Manually trigger a loop execution. ALWAYS runs the governance gate first — if governance denies or errors, the loop body does NOT execute. Returns the governance receipt and execution outcome. This is the core DDLP safety mechanism: declare intent, get governed, then act.
list_loops List all registered loops with their current status, schedule, risk tier, and last execution result. Use this to see what loops exist and their governance health.
pause_loop Pause a loop so it cannot be executed. Paused loops reject execute_loop calls until resumed. Use this to safely stop a loop without deleting its history.
resume_loop Resume a paused loop so it can be executed again. Clears the pause reason from metadata.
loop_audit Get the execution history for a specific loop. Shows every governance decision, outcome, and timestamp. Use this to audit how a loop has been behaving over time.
Permissions 3
network medium filesystem low env_vars low