openclaw-repair-sweep

Orchestrate worker fleets to diagnose, refactor, land, and close OpenClaw issues and pull requests.

388k|81.5k|Updated Nov 24, 2025
One-click install
npx skills add https://github.com/openclaw/openclaw --skill openclaw-repair-sweep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: openclaw-repair-sweep
Source: https://github.com/openclaw/openclaw/tree/main/.agents/skills/openclaw-repair-sweep
Command: npx skills add https://github.com/openclaw/openclaw --skill openclaw-repair-sweep

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Managing a large backlog of GitHub issues and pull requests across an active open-source project requires consistent triage, root-cause analysis, refactoring, and verified closure. This Skill automates that end-to-end campaign by coordinating isolated worker subagents over the issue queue.

Core Features & Use Cases

  • Fleet Orchestration: Spawns bounded worker subagents in isolated worktrees to investigate, fix, refactor, and land repairs in parallel.
  • Root-Cause Repair: Proves defects against current source and upstream dependencies before accepting a fix, preferring clean refactors over narrow patches.
  • Verified Closure: Traces ancestry, runs independent verifier challenges, and posts evidence-backed closure comments only when the primary symptom is resolved.
  • Use Case: Run a discovery sweep over the open issue queue to find the five best-qualified bugs, assign each to a worker, land the verified fixes, and close the proven items in one campaign.

Quick Start

Use the openclaw-repair-sweep skill to run a discovery batch of five repairs over the open OpenClaw issue queue and report landed PRs and closed issues.

Frequently Asked Questions about openclaw-repair-sweep

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate fixing a backlog of GitHub issues with AI workers?

Use a repair-sweep orchestration skill that spawns isolated worker subagents in per-issue worktrees. Each worker investigates, proves the root cause, implements a refactor, and lands a verified PR before the orchestrator closes the issue.

What is the best workflow for autonomous PR repair and merge?

Assign each issue or root-cause cluster to one bounded worker in an isolated worktree, require independent verifier challenges, and only land through the repo-native scripts/pr flow after CI is green on the exact pushed head.

Can AI agents safely merge pull requests without maintainer review?

Agents can land routine repairs when authorized, but product decisions, security changes, and public API boundary changes must be flagged as NEEDS-MAINTAINER-DECISION rather than merged automatically.

Why does autonomous issue closure fail without ancestry proof?

Closure requires git merge-base --is-ancestor verification that the fix SHA is reachable from current main and every affected build. Merge dates and similar wording are not sufficient evidence for closing an issue.

When should I prefer a refactor over a quick bug fix patch?

Prefer the coherent owner-boundary refactor when the minimal patch and the clean repair diverge, as long as the risk stays understood and bounded. Delete connected dead branches and duplicate paths in the same change.