Practical skills for AI coding agents to build features, debug issues, review code, and test changes.
-
Updated
Oct 4, 2026
Practical skills for AI coding agents to build features, debug issues, review code, and test changes.
Candid reviews of Claude Code skills. File:line citations. Ranked actions. No filler.
Evidence-led review and minimal refinement for existing agent skills | 既有 Agent Skill 的证据化审核与最小打磨
Your AI agent skill doctor - 5-dimension scoring + security gate + leaderboard
Security review skill for Claude Code. Scans third-party agent skills against 35 detection rules and 7 semantic checks before you install them. Finds patterns worth looking at, flags what could go wrong. Does not pass or fail anything.
Audit and score Codex skills with a transparent, evidence-based custom rubric
Free, open AI-agent skills from Majarrah Nexus — technical skill review and guided approval walkthroughs for Claude Code, Codex, Cursor, Copilot, and Gemini.
8-Dimensional Skill Analysis Engine for Claude Code. Forensic auditing with weighted scoring, automated pattern detection, and rich HTML reports. Zero dependencies.
8-Dimension Skill Review Tool for Hermes / Claude Code. Zero-dependency Python script with HTML report output.
🛡️ OpenClaw Skill Security Gatekeeper — Automated security auditing for AI agent skills. Detects prompt injection, data exfiltration, credential harvesting, and supply chain attacks.
Curated public Codex skills for diagrams, skill review, and agent workflows.
Codex skill review / skill 检验 tool for evaluating necessity, baseline lift, rubric score, and keep/simplify/merge/rewrite/delete/patch decisions.
Which AI agent Skills survive a real build? Every pick is field-tested by building a product from scratch, then rated on a public rubric - 15 published picks, 16 field tests, every demo live and open. by Phoebe Fu
Review and score an AI agent skill against an 11-dimension standard with automated hard checks. / 按 11 维标准审查并自动给智能体技能打分。
Claude Code skill that audits third-party skills for security before installation. Catches prompt injection, RCE, data exfiltration, and other attack patterns.
Security audits for Claude Code skills, written in plain English for non-technical builders and "vibe coders" — get a SAFE / CAUTION / UNSAFE verdict before you install anything.
To associate your repository with the skill-review topic, visit your repo's landing page and select "manage topics."