菜单
Query DeepWiki for repository documentation and structure. Use to understand open source projects, find API docs, and explore codebases.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Explore and compare architectural options before committing to a large technical direction. Use when the user wants to evaluate different architectures, compare approaches, choose between competing designs, rethink a subsystem, or understand tradeoffs before a major refactor or migration. Also use for prompts like "explore the architecture", "what are our options", "compare approaches", "what design should we choose", "audit and recommend an improved architecture", or "help me think through a large architectural change" even if the user does not mention a formal architecture review.
Build a compilable type-level skeleton from a high-level architecture spec before writing any implementation logic. Use when you have an architectural assessment, design doc, or restructuring plan and need to prove the new architecture is sound before migrating code. Also use when asked to "scaffold the new architecture", "create type stubs", "build the shell", "flesh out this spec", "skeleton the modules", or any request to turn architectural intent into verified structure. This skill follows the "Human Builds the Shell" paradigm: types are hard constraints that the compiler enforces, so if the skeleton compiles, the architecture is structurally sound. Especially valuable for large refactors where you don't trust agents to maintain coherence.
Give the human a fast, plain-English catch-up on what changed in the project: what the agents did, why, and what decisions need their input. Use this whenever the user asks to "catch me up", "what changed", "where are we", "recap", "brief me", "give me the rundown", "what did you do", "summarize the session", "fill me in", or otherwise signals they have been away and want to get back up to speed quickly. Built for someone steering several agent-driven projects at once who does not read the code closely but needs to grasp the core ideas, the choices made, and the open decisions well enough to steer. Trigger even if they do not use these exact words: any request to get oriented on recent progress should use this skill.
Analyze a provided resource through a Circuit lens. Use when the user provides or points to an article, pasted conversation, X thread, workflow example, GitHub repo, local codebase, product page, paper, or other outside reference and asks what Circuit can learn from it, whether it suggests a possible Circuit flow, whether Circuit can support the workflow today, what product or block changes would be needed, or whether the analysis should be added to Circuit's ideas or flow backlog.
Operate Circuit, the flow engine that runs coding work as structured, evidence-backed flows (Fix, Build, Explore, Review, Prototype). Use this skill at the START of any substantive coding task in a project where Circuit is installed: fixing a bug, building a feature, refactoring, reviewing a diff or PR, investigating a question, or prototyping. Also use it whenever the user mentions Circuit, flows, runs, checkpoints, run folders, the power dial or process override, connectors, or asks to create, generate, preview, configure, or resume a flow. If a coding request arrives and Circuit is available, consult this skill before doing the work by hand.
Forensic audit of the user's recent Claude Code sessions to surface step-change workflow improvements — not marginal ones. Use when the user asks to "audit my Claude Code sessions", "analyze how I use Claude Code", "find patterns in my usage", "improve my Claude Code workflow", "review my sessions", "find leverage in my setup", or wants to understand where their Claude Code setup is leaking time. Samples dozens of real transcripts, extracts quantitative signal via scripts, uses parallel subagents for deep reads, then synthesizes into a short prioritized report with drafted implementations (new skills, CLAUDE.md rules, hooks, settings diffs) that the user can install directly. Trigger even when the user doesn't say the word "audit" — if they're asking about improving or reviewing their Claude Code habits at scale, use this skill.
基于 SOC 职业分类
| name | deepwiki |
| description | Query DeepWiki for repository documentation and structure. Use to understand open source projects, find API docs, and explore codebases. |
Query DeepWiki for AI-generated documentation about any public GitHub repository.
DeepWiki (deepwiki.com) provides AI-generated documentation for GitHub repositories, including:
URL Pattern: Replace github.com with deepwiki.com in any repo URL:
github.com/vercel/next.js → deepwiki.com/vercel/next.jsDeepWiki provides a free MCP server with no authentication required for public repos.
claude mcp add -s user -t http deepwiki https://mcp.deepwiki.com/mcp
Add to your MCP config:
{
"mcpServers": {
"deepwiki": {
"serverUrl": "https://mcp.deepwiki.com/sse"
}
}
}
Once configured, these tools become available:
| Tool | Purpose |
|---|---|
read_wiki_structure | Get documentation topics/structure for a repo |
read_wiki_contents | Retrieve actual documentation content |
ask_question | Ask AI-powered questions about the repo |
# Fetch documentation overview
WebFetch https://deepwiki.com/owner/repo "Summarize the architecture"
# Example
WebFetch https://deepwiki.com/vercel/next.js "How does routing work?"
Use the MCP tools directly:
mcp__deepwiki__read_wiki_structure - Get repo structuremcp__deepwiki__read_wiki_contents - Get documentationmcp__deepwiki__ask_question - Ask questionsIf DeepWiki lacks coverage for a repo, use GitHub API:
gh api repos/owner/repo | jq '{description, language, topics, stars: .stargazers_count}'
gh api repos/owner/repo/readme --jq '.content' | base64 -d
gh api repos/owner/repo/git/trees/main?recursive=1 | \
jq -r '.tree[] | select(.type == "blob") | .path' | head -50
Two protocols are supported:
https://mcp.deepwiki.com/sse - Official MCP spechttps://mcp.deepwiki.com/mcp - Cloudflare/OpenAI compatible