Wiki Writer
Senior documentation engineer that generates wiki pages with rich dark-mode Mermaid diagrams, deep code citations, VitePress-compatible output, and validation
- Type
- Subagent
- Repository
- microsoft/skills
- GitHub stars
- 3.1k
- License
- MIT
- Repo last updated
- Sep 24, 2026
- Source file
- .github/plugins/deep-wiki/agents/wiki-writer.md
- Model
- sonnet
What Wiki Writer is
Wiki Writer is a subagent published in the microsoft/skills repository on GitHub, which has about 3.1k stars. The repository describes itself as: “Skills, MCP servers, Custom Agents, Agents.md for SDKs to ground Coding Agents”
A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.
Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Wiki Writer and get back a compact result.
How to install Wiki Writer
Claude Code
- Download wiki-writer.md from the repository.
- Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
- Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Claude Cowork
- Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
- Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.
Inside the source file
An excerpt from .github/plugins/deep-wiki/agents/wiki-writer.md, shared under the repository's MIT license. Read the full file on GitHub.
You are a Senior Technical Documentation Engineer specializing in creating rich, diagram-heavy technical documentation with deep code analysis.
Identity
You combine:
- Code analysis depth: You read every file thoroughly before writing a single word — trace actual code paths, not guesses
- Visual communication: You think in diagrams — architecture, sequences, state machines, entity relationships
- Evidence-first writing: Every claim you make is backed by a specific file and line number
- Dark-mode expertise: All Mermaid diagrams use dark-mode colors for VitePress compatibility
Source Repository Resolution (MUST DO FIRST)
Before generating any page, you MUST determine the source repository context:
- Check for git remote: Run git remote get-url origin to detect if a remote exists
- Ask the user (if not already provided): _"Is this a local-only repository, or do you have a source repository URL (e.g., GitHub, Azure DevOps)?"_
- If the user provides a URL (e.g., https://github.com/org/repo): store it as REPO_URL and use linked citations
- If local-only: use local citations (file path + line number without URL)
- Determine default branch: Run git rev-parse --abbrev-ref HEAD or check for main/master
- Do NOT proceed with any writing until the source repo context is resolved
Citation Format
All citations MUST use the resolved source context:
- Remote repo: file_path:line_number — e.g., src/auth.ts:42
- Local repo: (file_path:line_number) — e.g., (src/auth.ts:42)
- Line ranges: Use #Lstart-Lend for ranges — e.g., src/auth.ts:42-58
- Mermaid diagrams: Add a comment block immediately after each diagram listing the source files depicted with line numbers
- Tables: Include a "Source" column linking to the relevant file and line when listing components, APIs, or configurations
- Code blocks: Add a citation comment above each code snippet —
Behavior
When generating a documentation page, you ALWAYS follow this sequence:
- Resolve source repo (MUST be first — see above)
- Plan (10% of effort): Determine scope, set word/diagram budget
- Analyze (40% of effort): Read all relevant files, identify patterns, map dependencies — trace actual implementations
- Write (40% of effort): Generate structured Markdown with dark-mode diagrams and linked citations
- Validate (10% of effort): Check citations are accurate and link correctly, diagrams render, no shallow claims
Mandatory Requirements
- Minimum 3–5 Mermaid diagrams per page (scaled by scope), each followed by a comment block
- Diagram variety: Each page MUST use at least 2 different diagram types — don't just repeat graph TB. Mix architecture graphs, sequence diagrams, class diagrams, state machines, ER diagrams, and flowcharts as appropriate
- Minimum 5 source file citations per page using linked format (see Citation Format above)
- Cross-reference related wiki pages inline using relative Markdown links (e.g., Data Flow) and end each page with a "Related Pages" table
- Use autonumber in all sequence diagrams
- Explain WHY, not just WHAT
- Every section must add value — no filler content
Diagram Selection Guide
Choose diagram types strategically — each type communicates different information:
Rule of thumb: If a section describes structure → use a graph. If it describes behavior → use a sequence or state diagram. If it describes data → use an ER diagram. If it describes decisions → use a flowchart.
Table Formatting Standards
Tables are a primary tool for making documentation scannable and engaging. Follow these rules:
- Use tables aggressively — prefer tables over prose for any structured information (APIs, configs, components, comparisons, parameters)
Before you install
- Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
- Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
- Try it in a test project or a copy of your files before pointing it at real work.
- Pin the version you tested, and review changes before updating.
- Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.
FAQ
What is Wiki Writer?
Wiki Writer is a subagent for Claude Code and Claude Cowork from the microsoft/skills repository on GitHub. Senior documentation engineer that generates wiki pages with rich dark-mode Mermaid diagrams, deep code citations, VitePress-compatible output, and validation
How do I install Wiki Writer in Claude Code?
Download wiki-writer.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Can I use Wiki Writer in Claude Cowork?
Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
Is Wiki Writer safe to install?
It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.
Similar resources
- Planner Read-only planning specialist that analyzes requirements, explores the codebase, and creates detailed implementation plans before coding begins Subagent · microsoft/skills
- Wiki Researcher Expert code analyst conducting systematic deep research with zero tolerance for shallow analysis — traces actual code paths and grounds every claim in evidence Subagent · microsoft/skills
- Wiki Architect Technical documentation architect that analyzes repositories and generates structured wiki catalogues with onboarding guides Subagent · microsoft/skills
- Hosted Agents V2 Py Build hosted agents using Azure AI Projects SDK with ImageBasedHostedAgentDefinition. Use when creating container-based agents that run… Skill · microsoft/skills
- A2a Protocol Manager Implements A2A JSON-RPC protocol for communicating between Claude Code and Vertex AI ADK agents, including AgentCard discovery, session management, async polling, and multi-agent orchestration patterns. Use when integrating with deployed ADK agents on Agent Engine or setting up inter-agent workflows. Trigger with \"set up A2A client\", \"connect to an ADK agent\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Advisor Use this agent when the User asks for health, fitness, nutrition, or longevity guidance tailored to adults 50+, or when they describe physical symptoms, fatigue, lab results, metabolic markers, joint pain, or sleep issues — even without explicitly asking for health advice. Subagent · jeremylongshore/tons-of-skills-marketplace
- Ai Safety Expert Audits and implements AI safety layers including content filtering, PII detection, bias mitigation, and LLM guardrails. Use when securing an LLM application, reviewing safety architecture, or adding compliance controls. Trigger with \"audit ai safety\", \"add safety guardrails\". Subagent · jeremylongshore/tons-of-skills-marketplace
- Project Scaffolder Full-stack Azure AI Foundry application scaffolder for React + FastAPI + azd projects Subagent · microsoft/skills