# Agentlab Maintainer

> Run AgentLab quality gates and the fake-good engine self-test. Use when changing AgentLab itself.

- Skill: `ali975/agentlab-maintainer` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ali975/agentlab-maintainer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ali975/agentlab-maintainer/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ali975 (https://skillmd.com/u/ali975)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/ali975/agentlab-maintainer

---


# agentlab-maintainer

When working on this repository:

1. Run `npm run typecheck`, `npm run lint`, `npm test`, and `npm run build`.
2. After a CLI or engine change, run `npx agentlab test --agents fake-good`.
3. Do not treat `fake-good` as a native Codex/OpenCode/Claude result.
4. If you change a native adapter, add a regression fixture under `test/fixtures/native/` from sanitized `--help` or event JSON.
5. Never copy credentials into the workspace or commit `.agentlab/results/`.
6. Update `docs/COMPATIBILITY.md` only after a real native run, and never upgrade a NOT TESTED surface to PASS without evidence.