{
  "markdown": "<div align=\"center\">\n\n# 🧰 skill-kit\n\n### The skill-testing workshop, packaged as a Claude Code plugin\n\n**Author, evaluate, security-review, and autonomously improve Claude Agent Skills — in any project, with evidence, not vibes.**\n\n[![CI](https://github.com/mjenkinsx9/skill-kit/actions/workflows/harness.yml/badge.svg)](https://github.com/mjenkinsx9/skill-kit/actions/workflows/harness.yml)\n![Harness checks](https://img.shields.io/badge/harness_checks-21-blue)\n![Bash](https://img.shields.io/badge/bash-3.2%2B-4EAA25?logo=gnubash&logoColor=white)\n![Python](https://img.shields.io/badge/python-3.x-3776AB?logo=python&logoColor=white)\n![Platform](https://img.shields.io/badge/platform-linux%20%7C%20macos%20%7C%20windows-lightgrey)\n![Made for Claude Code](https://img.shields.io/badge/made_for-Claude_Code-d97757)\n[![License: MIT](https://img.shields.io/badge/license-MIT-yellow.svg)](LICENSE)\n\n</div>\n\n---\n\nThe core idea: **a skill is only as good as the evidence behind it.** skill-kit\nbrings the full evaluation stack from the\n[skill-testing](https://github.com/mjenkinsx9/skill-testing) dev-bench to any\nproject as an installable plugin — a 21-check static harness, a live behavioral\ncheck, empirical trigger measurement, a blind value-add baseline, and an\nautoresearch loop that iteratively tightens a `SKILL.md` and keeps only the\nchanges that score better.\n\n## ✨ What's inside\n\n| Layer | Tool | What it proves |\n|---|---|---|\n| 🧹 **Static** | `check-skill` | The skill is *well-formed* — 21 checks: frontmatter, naming, trigger language, secret & security-smell scans |\n| ⚡ **Behavioral** | `behavioral-check` | The documented commands *actually run* against the real system today |\n| 🎯 **Empirical trigger** | `trigger-accuracy` | The skill *really fires* on its positive prompts and declines its negatives |\n| ⚖️ **Value-add** | `value-add-test` | The skill *beats the cold model* in a blind head-to-head |\n| 🔄 **Autoresearch** | `improving-skills` | Iterative modify → score → keep-or-revert loop that tightens an existing SKILL.md |\n| ⚓ **Goal anchoring** | `goal-new-skill` · `goal-improve-skill` | Skill work is driven by a measurable end state, not vibes |\n\nThe harness is self-tested (9 pathological fixtures + 34 pytest tests) and CI\nenforces zero FAILs *and* zero WARNs on the shipped skills, on every push and PR.\n\n## 🚀 Install\n\nInside Claude Code, on any machine:\n\n```\n/plugin marketplace add mjenkinsx9/mjenkins-toolbox\n/plugin install skill-kit@mjenkins-toolbox\n```\n\nUpdate later with `/plugin update skill-kit`. Because a plugin's `bin/` joins the\nBash `PATH`, every helper is callable by bare name in any repo.\nskill-kit also installs on Copilot CLI, Codex, Cursor, and Gemini —\nsee [Multi-harness support](docs/04-harnesses.md).\n\n```bash\n# Lint a skill directory (or a SKILL.md directly) — exit 0 iff zero FAILs:\ncheck-skill .claude/skills/my-skill\n```\n\n```\n/skill-kit:improving-skills .claude/skills/my-skill/SKILL.md\n```\n\n## 📚 Documentation\n\nFull guides live in [`docs/`](docs/). Start with the overview, then dive into the\nsection you need.\n\n| Doc | Description |\n|---|---|\n| [Overview](docs/01-overview.md) | What skill-kit is, the core idea, and the evaluation stack |\n| [Installation](docs/02-installation.md) | Marketplace install, the `bin/`-on-`PATH` convention, requirements |\n| [Usage](docs/03-usage.md) | The `bin/` commands, slash commands, and the improving-skills scoring loop |\n| [Multi-harness support](docs/04-harnesses.md) | Per-harness status, manifests, and install notes |\n| [Repository layout](docs/05-layout.md) | Map of manifests, helpers, commands, skills, and tests |\n| [Testing the tests](docs/06-testing.md) | Self-tests + pytest, and the relationship to skill-testing |\n\nFull documentation map: [docs/README.md](docs/README.md)\n\n## 📄 License\n\n[MIT](LICENSE) © 2026 Mike Jenkins · [Security policy](SECURITY.md)\n",
  "bytes": 3872,
  "sha": "777ae9518249acab8e6663a66df9ab349bd6b3ebd94be556b47289d581e0f80e",
  "repo_slug": "crashcartlabs/skill-kit",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_crashcartlabs_skill_kit_ae1ddf32/readme"
}