skill-regressionlisted

Regression testing framework for AgentSkills. Analyzes a target skill, runs script-layer assertions and AI-layer semantic scoring, and outputs a Markdown report. Supports two backends: OpenAI-compatible LLM (default, works anywhere) or OpenClaw cron-based agent trigger (auto-detected if openclaw CLI present). Use when the user asks to "test a skill", "regression test xxx skill", "run skill QA", or "audit my skill against test cases". Requires SR_LLM_API_KEY env var or interactive setup; supports .env files.
Songhonglei/build-better-skills · ★ 1 · AI & Automation · score 74

Install: claude install-skill Songhonglei/build-better-skills

# Skill Regression - **Version**: 1.0.3 - **License**: MIT - **Author**: Evan Song · [github.com/Songhonglei](https://github.com/Songhonglei) - **Repository**: https://github.com/Songhonglei/build-better-skills/tree/main/skills/skill-regression - **Part of**: [`build-better-skills`](https://github.com/Songhonglei/build-better-skills) suite — see [Stages](https://github.com/Songhonglei/build-better-skills#stages) for the lifecycle map. ## What it does Given a target skill directory, this skill: 1. Normalizes test file naming (TEST.md is canonical; legacy Test.md/TESTS.md auto-renamed) 2. Loads or LLM-generates structured test cases 3. Runs script-layer assertions (subprocess + regex/contains/exact) 4. Runs AI-layer semantic scoring (LLM-as-judge, configurable threshold) 5. Outputs a Markdown report (with optional upload hook) Two execution backends, auto-detected: | Backend | When | What it tests | |---------|------|---------------| | `api` (default outside OpenClaw) | LLM API + system prompt | Whether SKILL.md is well-written enough for LLMs to follow correctly | | `openclaw` (default inside OpenClaw) | Cron probe job → real agent → jsonl polling | End-to-end agent + skill integration | ## Quick start ```bash # 1. First time: configure your LLM python3 scripts/setup.py # 2. Run regression on a skill bash scripts/run_regression.sh /path/to/my-skill # 3. Detailed mode (full input/output per case) bash scripts/run_regression.sh /path/to/my-skill --detail # 4. Re-run o