Every skill & agent
for Claude Code
A marketplace for Claude Code extensions: 25,757 skills and 13,833 agents crawled from public GitHub, each shown with its source, the tools it touches, and a working install command. Free to browse, free to be listed.
What is Cult of Claude?
Cult of Claude is an independent marketplace and directory of Claude Code skills and agents. A crawler
visits public GitHub every couple of hours, finds SKILL.md files, agent definitions and
slash-command files, scores each one, and publishes an entry page for anything that clears the quality gate.
Nothing here is hosted by us: every entry links to the repository it came from, and installing means
copying that author's file onto your machine. The site is run by FDF Labs and is
not affiliated with Anthropic.
What is a skill? What is an agent?
A skill is a folder whose SKILL.md teaches Claude Code a procedure: how to
review a pull request, write a migration, or drive a testing tool. Claude loads it when you type its name
or when its description matches what you asked for. An agent (a subagent) is a Markdown
file that defines a separate Claude with its own system prompt, tool list and context window, used for
work that would otherwise flood your conversation. Plugins bundle both, hooks run shell on lifecycle
events, and MCP servers add tools; the
comparison guide sorts them out.
How does installation work?
There is no package manager for a plain skill. You copy the folder into ~/.claude/skills/
for every project, or into a repository's .claude/skills/ for one team; agents go to
~/.claude/agents/ as a single file. Each entry page prints the exact command for its source
path, with a clone-and-copy for skill folders and a curl for single files, and says which
directory it writes to. Read the file before you run the command: a skill can pre-approve tools and attach
hooks. The install guide covers project-level
installs, plugin marketplaces, verifying the skill loaded, and removing it.
New This Week
Entries first seen by the crawler in the last seven days, ranked by score. The full feed of new entries is in the Atom feed.
Discover
A daily rotation of well-documented community entries scoring above 50, one per source repository, so the same six repos do not sit here forever.
root-cause-tracing
80Use when errors occur deep in execution and you need to trace back to find the original trigger.
System Cache Cleanup
72Clean temporary files and caches to free disk space: $ARGUMENTS
Ecc Guide
75Navigate ECC's current agents, skills, commands, hooks, install profiles, and docs from the live repository surface.
Vercel Deploy
70- Deploy applications and websites to Vercel with preview or production options
Skill Development Guide
79[Community Examples](https://github.com/anthropics/claude-code/tree/main/examples)
impeccable
70| The design language that makes your AI harness better at design. | 13448 | 27 | 1 |
docs-architect
74| You are a technical documentation architect specializing in creating comprehensive, long-form docume... | sonnet | [wshobson/agents](https://github.com/wshobson/agents) |
Chatgpt On Wechat
76CowAgent是基于大模型的超级AI助理,能主动思考和任务规划、访问操作系统和外部资源、创造和执行Skills、拥有长期记忆并不断成长,比OpenClaw更轻量和便捷。同时支持微信、飞书、钉钉、企微、QQ、公众号、网页等接入,可选择OpenAI/Claude/Gemini/DeepSeek/ Qwen/GLM/Kimi/LinkAI,能处理文本、语音、图片和文件,可快速搭建个人AI助理和企业数字员工。
Ccxt Pr Reviewer
66End-to-end review of a CCXT pull request. Reads the diff, transpiles and builds in all six languages (TS/JS, Python, PHP, C#, Go, Java), runs offline tests + live smoke tests, inspects both source and
Gitnexus Branch Hygiene Reviewer
66GitNexus branch hygiene and mergeability reviewer. Use to classify merge state, conflicts, stale branches, merge-from-main commits, unrelated churn, mixed domains, and whether rebase or split is requi
Context Query Agent
63You are a specialized agent for querying the Artifact Index to find relevant precedent.
Resource Duplication
65Resource duplication expert - use when configuring a resource to be copied between projects.
Feeling Lucky?
Spin the wheel to discover something new
Pick a category
You discovered...
Browse skills by category
Every skill is filed under one of 12 categories. Each hub lists the category's best-scoring entries first, with the full filtered grid below.
- Development 12,062
- AI 4,612
- Git 1,312
- Security 1,252
- Testing 1,230
- Design 1,131
- Data 1,076
- Productivity 1,061
- Documentation 691
- DevOps 590
- Communication 526
- Code Quality 209
Agents use their own taxonomy of 11 categories; browse them on the agents page, or see the full breakdown on the stats page.
Top Rated Skills
View allThe score beside each entry is out of 100 and is computed, not voted. Twelve points per order of magnitude of GitHub stars, then fixed bonuses for a README, a long description, a trigger condition, a declared tool list and an official Anthropic origin. It rewards documentation and adoption, not correctness, and it is never a security review. Before any entry reaches this page it also had to clear a separate quality gate that removes bulk-generated templates, MCP servers mislabelled as skills, and entries with no content behind them. The full formula is on the about page.