autoreview

Originally fromopenclaw/openclaw
Installation
SKILL.md

Auto Review

Run the bundled structured review helper as a closeout check. This is code review, not Guardian auto_review approval routing.

Rules

  • Start only with authoritative task context, completed builder guardrails, and real-surface proof; cite the evidence or report the missing prerequisite.
  • Findings are advisory. Verify before fixing, stay inside the task scope, and do not make a ship decision or invoke additional reviewer workflows.
  • Honor the requested engine and model. Use panels only when explicitly requested or justified by risk.
  • Treat the validated bundle as the reviewer's only repository input. Never bypass its required secret scan, sensitive-path omissions, or fail-closed input guards.
  • Split refused oversized bundles into dependency-connected targets, verify cross-target contracts directly, and never claim whole-change cleanliness from independent chunk reviews.
  • If the source tree changes after bundle creation, discard the result and rerun against the updated tree.
  • Do not push just to review. Push only when the user requested push, ship, or PR update.
Installs
88
GitHub Stars
4
First Seen
May 23, 2026
autoreview — uinaf/agent-skills