[Milestone v3][P1] Full-auto dry-run E2E with guardrails #49
Description
Summary
Run a full autonomous workflow dry-run from issue intake to PR review iteration with guardrails enabled (without auto-merge).
Why
We now have loop control-plane, policy gates, observability, and OpenClaw v2 command registry. We need an end-to-end proof run that validates behavior under realistic conditions.
Scope
- Trigger workflow from a test issue event
- Execute autonomous run through code change + PR creation path
- Run review-iteration once (poll checks/comments, apply one follow-up if needed)
- Keep merge manual for this dry-run
- Capture timeline and artifacts
Acceptance Criteria
- Dry-run executes from trigger to merge-ready PR without manual code edits outside defined intervention points.
- Guardrails/policy decisions are visible in run/event logs.
- Deterministic failure handling is verified for at least one synthetic invalid-input case.
- Report added under
.agent/reports/execution-reports/with links to issue/PR/checks.
Priority
P1 (highest)