skills.rest
openclaw
openclawOFFICIAL
Explore all skills by this author →
Website𝕏@openclawGitHub

auto-qa

Continuously audit, reproduce, and merge root-cause fixes across OpenClaw subsystem lanes.

388k|81.5k|Updated Nov 24, 2025
Search Tags:#root cause#QA#live testing#Openclaw#Code review#Pull request#subsystem lanes
One-click install
npx skills add https://github.com/openclaw/openclaw --skill auto-qa
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: auto-qa
Source: https://github.com/openclaw/openclaw/tree/main/.agents/skills/auto-qa
Command: npx skills add https://github.com/openclaw/openclaw --skill auto-qa

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

OpenClaw is a large, multi-subsystem codebase where bugs hide across gateway, CLI, UI, plugins, native apps, and packaging surfaces. This Skill runs a continuous, current-main QA campaign that reproduces real defects, deduplicates them against the ledger, and lands verified root-cause refactors instead of symptom-masking patches.

Core Features & Use Cases

  • Multi-lane subsystem auditing: Maintains at least ten independently scoped investigations covering CLI, gateway HTTP/transport, OpenAI provider, plugins, Control UI, QA Lab, agent sessions, schedulers, native apps, channels, and packaging.
  • Live product proof: Routes real provider, gateway, browser, packaged-install, and soak tests through isolated state and ports, requiring actual model responses, nonzero scenario counts, and rendered UI screenshots.
  • Root-cause refactor workflow: Reproduces the failing user path, repairs the canonical owner and all affected siblings, adds authentic regression coverage, runs independent review, and merges only authorized low-risk fixes.
  • Use Case: An operator wants 100 verified bug fixes landed on main. Activate this Skill to spin up parallel reviewer lanes, reproduce each defect against frozen SHAs, refactor the canonical owner, and autonomously merge only the low-risk, independently reviewed root-cause repairs.

Quick Start

Run the auto-qa skill against the current OpenClaw main branch to audit at least ten subsystem lanes, reproduce real bugs, and land verified root-cause fixes.

Frequently Asked Questions about auto-qa

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

FAQPage Schema
How do I run an autonomous QA campaign on the OpenClaw codebase?▼

Activate the auto-qa skill against the current main branch. It freezes the canonical SHA, spins up at least ten subsystem reviewer lanes, reproduces real defects, and merges verified root-cause fixes into a numbered ledger.

What subsystems does the OpenClaw auto-qa skill cover?▼

It covers CLI setup, gateway HTTP and transport, OpenAI provider, plugin lifecycle, Control UI, QA Lab scenarios, agent sessions, schedulers, native and portability, channel adapters, and packaging and distribution.

Can auto-qa autonomously merge pull requests without review?▼

Only low-risk root-cause refactors that are individually reproduced, independently reviewed, current-main-compatible, and have green required exact-head proof. Security, migration, protocol, SDK, and architectural changes are marked user-review-required and excluded from the count.

Why does auto-qa reject symptom-only patches and minimal diffs?▼

Symptom-masking guards, compatibility shims, and narrowly passing tests leave the architectural defect intact. The skill requires repairing the canonical owner and all affected siblings so the shared invariant becomes obvious and reliable.

How does auto-qa verify live provider behavior instead of mocks?▼

It inspects the real configured agents.list, resolves an existing agent, requires a canonical openai/<model> selection, and proves both the delivered model-final path and the separately persisted transcript or session as distinct evidence.

When should I not use the auto-qa skill?▼

Avoid it for security-sensitive, persistent-state, protocol, public-contract, or plugin SDK changes that require maintainer authorization, and for offline work where remote freshness cannot be verified.

Related Skills

Explore More Software Engineering Skills→
Apostolos-DelisApostolos-Delis
1

rebase

Rebase a Git branch onto a base branch with conflict resolution and safe pushing.

Community
Advanced
andreciniandrecini
2

write-changelog-entry

Generate a new CHANGELOG.md release entry from merged pull requests and commits.

Community
Intermediate
lionheartapplionheartapp

performance-testing

Run k6 and Locust load, stress, spike, and soak tests against APIs.

Community
Intermediate
JKHeadleyJKHeadley
75

iterative-converging-audit

Repeatedly audit a full surface until a sweep returns zero new findings.

Community
Advanced