What kind of visualizations does the skill provide?

Wiki Researcher generates multiple Mermaid diagrams per session, including architecture maps, sequence diagrams for data flow, and state diagrams for logic transitions.

How does Wiki Researcher ensure accuracy in its reports?

The skill follows a strict 'zero tolerance for shallow research' policy, requiring actual code path tracing and specific file/line evidence for every claim made.

Is this tool effective for legacy codebases with no documentation?

Yes, it is specifically designed for 'black box' investigation, building its own internal map of the system by following imports and call chains from the ground up.

Can it link directly to my GitHub or Azure DevOps repository?

Yes, it automatically detects your git remote context and formats citations as direct links to your specific repository, branch, and line numbers.

Codebase Deep Researcher

Name: Codebase Deep Researcher
Author: microsoft

bymicrosoft

•

1,777

•

학습 및 문서화

Conducts multi-turn, iterative deep research into codebases to uncover complex architectural patterns and data flows with evidence-backed analysis.

Wiki Researcher is a rigorous diagnostic tool designed for developers who need more than a superficial summary of their source code. It performs a structured five-step iterative analysis—covering architectural structure, data flow, integration dependencies, design patterns, and final synthesis. By resolving source repository contexts and enforcing a strict policy against 'vibes-based' analysis, it ensures every claim is grounded in actual code paths. The skill automatically generates Mermaid diagrams and verified file citations, making it an essential companion for onboarding to large projects, auditing legacy systems, or investigating complex bugs across distributed files.

주요 기능

01Automatic source repository resolution for hyperlinked GitHub/DevOps file citations.

02Evidence-based grounding that explicitly distinguishes between verified facts and inferences.

03Mandatory Mermaid diagram generation for visualizing architecture and sequence flows.

04Iterative 5-stage deep analysis covering structure, data flow, dependencies, patterns, and synthesis.

05Comprehensive technical debt identification and design pattern cataloging.

061,777 GitHub stars

사용 사례

01Tracing complex data flows and state management across multiple files to identify bottlenecks.

02Onboarding to a large, unfamiliar codebase to quickly understand system-wide interactions.

03Performing architectural audits to map dependencies and identify technical debt or anti-patterns.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microsoft/skills wiki-researcher

For use in Claude.ai and ChatGPT

Download Skill