Awesome Agent Skills

Skills for AI coding agents: code review, test writing, design docs, debugging, security audits, refactoring, and cleaning up AI-written code and text. Each skill is a folder with a SKILL.md in the Agent Skills format, distilled from public collections and large open-source practice, then sharpened on real repositories.

Works with Claude Code, Claude.ai, OpenAI Codex, Gemini CLI, Cursor, GitHub Copilot, opencode, Amp, Windsurf, Antigravity, Kilo Code, Qwen Code, and any other agent that supports the standard.

• Awesome Agent Skills
  • Contents
  • Quick start
  • Install as a plugin
  • Skills
    • Code review
    • Code quality and refactoring
    • Testing
    • Design and planning
    • Understanding a codebase
    • Debugging and reliability
    • Audits
    • Git and repository operations
    • Agent maintenance
    • Writing and text
    • Content marketing
  • Usage examples
  • Skill dependencies
  • Skill format
  • Contributing
  • Related
  • License

git clone https://github.com/khasky/awesome-agent-skills.git
cd awesome-agent-skills
npx skills add ./skills -g -s "*" -a claude-code codex gemini-cli -y

The skills CLI installs every skill globally into Claude Code, Codex, and Gemini CLI. Change -a to pick agents (-a "*" installs to every detected agent). To skip the clone, npx skills add khasky/awesome-agent-skills installs straight from GitHub.

Update with git pull in the clone, or npx skills update -g if the CLI copied instead of linking (check ls -l ~/.agents/skills: an entry with no link target is a copy). Re-run skills add to pick up new skills.

The plugin installs every skill at once and updates with the repository. Pick the plugin or the Quick start, not both.

Qwen Code prefixes a plugin skill with the extension name: awesome-agent-skills:awesome-code-review. CI installs the plugin into Claude Code, Codex, Copilot CLI and Gemini CLI on every push; the other rows follow each agent's documentation.

All audits are read-only and end in a SHIP / FIX / BLOCK verdict where they gate a release.

Skills trigger on intent, so plain requests work. Naming a skill makes the choice deterministic:

• "Use awesome-code-review on the diff against main."
• "awesome-bug-fix for this error: reproduce first, root cause before any fix — [paste logs]."
• "Audit this repo with awesome-leak-audit before we open-source it."
• "awesome-humanize-en: remove the AI voice from this draft."

They also chain: implement → awesome-code-cleanup → awesome-code-review before merge. Or: awesome-content-campaign → review the drafts → awesome-content-publisher.

The four skills that rewrite a published history or post to live accounts (awesome-git-history-reset, awesome-git-history-rebuild, awesome-git-author-rewrite, awesome-content-publisher) run only when you name them explicitly, never on a matching request alone.

Some skills read a reference file that ships with another skill instead of keeping their own copy, so one vocabulary or platform table has one source. The plugin and the Quick start install every skill, and nothing breaks. Installing a single skill (npx skills add ... -s <name>) leaves these links dangling: install the skills in its row too.

scripts/lint.py fails when this table and the links in the skills disagree.

Each skill follows the Agent Skills specification:

skills/<skill-name>/
├── SKILL.md          # required: YAML frontmatter (name, description) + instructions
└── references/       # optional: detailed docs the agent loads on demand

A skill is Markdown and nothing else: no scripts, templates or fixtures, so nothing here runs the author's code on your machine. The frontmatter description tells the agent when to activate; the body loads only on activation, and references/ only when needed.

Reports share two vocabularies: findings are rated Critical / High / Medium / Low / Informational, and release-gating audits end in SHIP / FIX / BLOCK, paired with NOT ASSESSED for what could not be checked.

One gate, the same locally and in CI:

python3 scripts/lint.py          # run the gates by hand
python3 scripts/test_lint.py     # prove each gate fails on a broken copy
python3 scripts/install-hooks.py # once per clone: run them before every commit

Each skill went through three passes: a survey of public collections and documented practice (kernel patch rules, OWASP, WCAG, PTES, NIST SP 800-115), a distillation that keeps only what changes an agent's output, and rounds of use on real repositories. Skills are self-contained, portable, stack-agnostic, and end with the check that proves the claim.

• Awesome AGENTS.md — the base, tool-agnostic ruleset every agent imports. Rules keep everyday work in line; skills here handle the jobs you name. Install both.
• Agent MCP Integrations — MCP servers that connect agents to browsers, cloud, databases, infra, and domain APIs.
• Claude Code Token Optimization — the token-efficiency layer.
• Claude Code Security Audit — the layered security-audit workflow.

Released under the MIT license. The awesome-humanize-en skill adapts material from humanizer-ru and, for its venue rules, editing-trace tests and evidence ledger, from sepia; awesome-seo-audit cross-checked its Google-policy detail against claude-seo, from which it also took the falsifiability line on every recommendation, the fix-ordering rule and the refusal of unsafe fetch targets, and took its collection mechanics from on-page-seo; its two-axis verdict follows claude-seo-ai and its evidence-and-confidence contract follows Agentic-SEO-Skill. awesome-agents-md-generator takes its section set and its differences-only nested files from agentseed, and its file format from agents.md. awesome-code-explainer takes its signals-first method, its input forms and its reading-path and open-questions sections from ExplainThisRepo. awesome-code-docs takes its dependency-graph method — symbol resolution, topological order, bottom-up summaries — from AutoDocs.