Deep Wiki: Single Page Generation

Generate a comprehensive wiki page for the specified topic.

Source Repository Resolution (MUST DO FIRST)

Before generating any page, resolve the source repository context:

1. Check for git remote: Run git remote get-url origin
2. Ask the user: "Is this a local-only repository, or do you have a source repository URL?"
   • Remote URL → store as REPO_URL, use linked citations: [file:line](https://github.com/microsoft/skills/blob/cf77b1efbf3117501f4727c476894751311ee885/.github/plugins/deep-wiki/commands/REPO_URL/blob/BRANCH/file/README.md#Lline)
   • Local → use (file_path:line_number)
3. Determine default branch: Run git rev-parse --abbrev-ref HEAD
4. Do NOT proceed until resolved

Inputs

The user will provide a topic/title and optionally specific file paths. Use $ARGUMENTS to determine what to document.

Depth Requirements (NON-NEGOTIABLE)

1. TRACE ACTUAL CODE PATHS — Do not guess from file names. Read the implementation. If function A calls B calls C, follow it all the way.
2. EVERY CLAIM NEEDS A SOURCE — File path + function/class name. "X calls Y" must include where.
3. DISTINGUISH FACT FROM INFERENCE — If you read the code, say so. If inferring, mark it explicitly.
4. FIRST PRINCIPLES — Explain WHY something exists before explaining what it does.
5. NO HAND-WAVING — Don't say "this likely handles..." — read the code and state what it ACTUALLY does.

Mandatory Three-Phase Process

Phase 1: Strategic Planning (ALWAYS FIRST)

1. Clarify the page's goals, audience, and deliverables
2. Determine scope based on relevant file count:
   • ≤50 files: full coverage
   • 50–300 files: prioritize critical paths

   • 300 files: tiered sampling (entry points, domain models, data access, integration edges)

3. Set documentation budget:
   • Small scope: ~2,000–3,000 words, 3 diagrams (2+ types)
   • Medium scope: ~3,000–5,000 words, 4 diagrams (3+ types)
   • Large/Complex: ~5,000–8,000+ words, 5–8 diagrams (4+ types)

Phase 2: Deep Code Analysis

1. Read ALL relevant source files completely
2. Identify: architecture patterns, design patterns, algorithms, data flow, state management
3. Map: component dependencies, external integrations, API contracts
4. Record citation anchors: file_path:line_number for every claim

Phase 3: Document Generation

Structure the page with:

• VitePress frontmatter: title and description
• Overview: purpose, scope, executive summary — explain WHY this exists
• At-a-glance summary table: Key components/concepts with one-line descriptions and source links — readers should grasp the system in 30 seconds
• Architecture / System Design: with graph TB/LR Mermaid diagram
• Core Components: purpose, implementation, design patterns — use a table per component group with "Component", "Responsibility", "Key File", "Source" columns
• Data Flow / Interactions: with sequenceDiagram (use autonumber)
• State / Lifecycle: with stateDiagram-v2 if the system has meaningful state transitions
• Data Model: with erDiagram if the system has entities or database tables
• Implementation Details: key algorithms, error handling, state management
• Configuration & Deployment: use tables for config options (Key, Default, Description, Source)
• References: inline citations throughout using resolved format
• Cross-references: Link to related wiki pages using relative Markdown links (e.g., [Data Flow](https://github.com/microsoft/skills/blob/cf77b1efbf3117501f4727c476894751311ee885/.github/plugins/deep-wiki/02-architecture/data-flow.md)). Whenever a concept, component, or pattern is covered in more depth on another wiki page, link to it inline. Also add a "Related Pages" section at the end listing connected wiki pages.

Content Organization Rules

• Progressive disclosure: Big picture first → drill into specifics. Don't front-load implementation details.
• Distill, don't dump: Every paragraph should earn its place. If a section is just listing things, convert it to a table.
• Tables over prose: For any structured data (APIs, parameters, configs, components, comparisons), ALWAYS use a table.
• One idea per paragraph: Keep paragraphs focused and scannable. Use bold for key terms.
• Section summaries: Start complex sections with a 1-2 sentence TL;DR before the details.
• Visual rhythm: Alternate between prose, diagrams, tables, and code blocks — avoid long walls of text.

Mermaid Requirements

Include minimum 3 diagrams using at least 2 different types. More is better — aim for one diagram per major section:

Type	Best For	When to Use
graph TB/LR	Architecture, component relationships	Structural overviews, dependency graphs
sequenceDiagram	API flows, interactions (always use autonumber)	Multi-step processes, request lifecycles
classDiagram	Class hierarchies, interfaces	Domain models, type relationships
stateDiagram-v2	State machines, lifecycle	Status transitions, workflow states
erDiagram	Database schema, entities	Data models, table relationships
flowchart	Data pipelines, decision trees	Conditional logic, error handling paths

Dark-mode colors (MANDATORY):

• Node fills: #2d333b, borders: #6d5dfc, text: #e6edf3
• Subgraph backgrounds: #161b22, borders: #30363d
• Lines: #8b949e
• If using inline style directives, use dark fills with ,color:#e6edf3
• Do NOT use <br/> in labels (use <br> or line breaks)