quality-validation — Claude Code skill: what it does, how to install
quality-validation blocks AI agents from claiming a task is finished without running fresh verification commands and checking exit codes.
LicenseRef-Proprietary 142 installs ★ 50 updated 2026-01-05
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a claude-codeWhat quality-validation does
This skill enforces an evidence-first policy for coding assistants. Before declaring any task finished, the model must follow a five-step gate procedure: choose the right command, execute it completely, inspect the output and status code, and confirm actual success. Guesses and assumptions are forbidden.
The rule set maps explicit verification requirements to specific claims. Passing tests require zero reported failures on a fresh run, successful builds require an exit code of zero, and bug fixes require executing the original failure case. When an agent delegates work to a subagent, it cannot trust the report and must inspect the version control diff directly.
Without a fresh terminal run in the active turn, the assistant cannot declare success. The skill ships under a proprietary license from the author.
What it helps with
- stop premature task completion claims from coding assistants
- require zero-failure test runs before staging commits or PRs
- ensure regression tests fail before fixes are restored
- verify subagent changes by checking repository diffs directly
Example prompts for the agent
- Fix the auth token expiry bug and confirm it with a fresh test run
- Review our requirements checklist and verify every item before committing
- Write a regression test for issue #88 and verify both the red and green stages
How to install quality-validation
1Skills CLI (Vercel)
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a claude-code2Manually: put the files in place
~/.claude/skills/quality-validation/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/quality-validation/
1Skills CLI (Vercel)
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a codex2Manually: put the files in place
.agents/skills/quality-validation/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a cursor2Manually: put the files in place
.cursor/skills/quality-validation/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install qodex-ai/ai-agent-skills skills/quality-validation2Manually: put the files in place
.github/skills/quality-validation/SKILL.md1Skills CLI (Vercel)
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a gemini-cli2Platform CLI
gemini skills install https://github.com/qodex-ai/ai-agent-skills --path skills/quality-validationManually: put the files in place
.devin/skills/quality-validation/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about quality-validation
How do I install quality-validation in Claude Code?
Run npx skills add qodex-ai/ai-agent-skills --skill quality-validation -a claude-code. The skill file will be stored in ~/.claude/skills/quality-validation/SKILL.md.
What does the quality-validation skill do?
It stops the agent from guessing or relying on previous runs. Every claim that tests pass, code compiles, or a bug is resolved requires executing a complete verification command and inspecting its exit status.
How does it prevent unverified subagent work?
The skill rejects blind trust in delegated agent reports. It directs the main assistant to inspect the version control diff and confirm file modifications independently.