| name | deepwiki |
| description | Quick AI-generated overview of GitHub repositories via DeepWiki MCP. Use when needing a fast high-level understanding of a repository's architecture and design decisions without fetching full repository.
|
| allowed-tools | ["Bash(*mcpcall.py*:*)"] |
DeepWiki
AI-powered documentation for GitHub repositories. Browse auto-generated docs, explore topic structures, and ask questions about any public repo. Powered by DeepWiki MCP. No API key required.
read_wiki_structure
Get a list of documentation topics for a repository. Use this first to discover what's documented.
repoName (required): GitHub repo in owner/repo format
scripts/mcpcall.py read_wiki_structure repoName:"facebook/react"
scripts/mcpcall.py read_wiki_structure repoName:"anthropics/claude-code"
read_wiki_contents
View full documentation about a repository.
repoName (required): GitHub repo in owner/repo format
scripts/mcpcall.py read_wiki_contents repoName:"facebook/react"
scripts/mcpcall.py read_wiki_contents repoName:"pallets/flask"
ask_question
Ask any question about a repository and get an AI-powered, context-grounded response.
repoName (required): GitHub repo (owner/repo string) or array of up to 10 repos
question (required): the question to ask
Single repo
scripts/mcpcall.py ask_question repoName:"facebook/react" question:"How does the reconciliation algorithm work?"
scripts/mcpcall.py ask_question repoName:"pallets/flask" question:"How are blueprints registered?"
Cross-repo comparison (up to 10 repos)
scripts/mcpcall.py ask_question --args '{"repoName": ["pallets/flask", "django/django"], "question": "How do these frameworks handle middleware?"}'
Tool Selection Guide
| Scenario | Tool |
|---|
| Discover what topics are documented | read_wiki_structure |
| Read full repo documentation | read_wiki_contents |
| Ask a specific question about a repo | ask_question |
| Compare multiple repos | ask_question with array of repos |
Tips
- Start with
read_wiki_structure to see available topics before diving into full contents.
ask_question supports up to 10 repos at once for cross-project comparisons.
- Use
ask_question for targeted queries — it's faster than reading full wiki contents.
- Works with any public GitHub repository.
- Responses are AI-generated and can hallucinate, especially on small/obscure repos. Always verify key claims against actual source code (use
/repo-cache to clone and read).