Generate Agent Documentation
Creates user-friendly documentation for AI coding assistant customization files.
Provider Folder Reference
This skill works across multiple AI coding assistant providers:
Provider Base Folder
GitHub Copilot .github/
Claude Code .claude/
Codex .codex/
OpenCode .config/opencode/
Throughout this document, <provider>/ represents your chosen provider's base folder.
When to Use
-
Onboarding new team members to your agent ecosystem
-
Creating a catalog of available agents and skills
-
Generating usage examples for specific agents
-
Documenting your customization setup for reference
Documentation Types
Individual Item Documentation
Generate detailed docs for a single agent, skill, prompt, or instruction.
Catalog Documentation
Generate a comprehensive overview of all customization files.
Output Templates
Agent Documentation
Agent: [Agent Name]
Overview
[What this agent does and why it exists]
When to Use This Agent
Use @[agent-name] when:
- [Scenario 1]
- [Scenario 2]
- [Scenario 3]
Don't use when:
- [Anti-pattern 1]
- [Anti-pattern 2]
How It Behaves
This agent will:
- [Behavior 1]
- [Behavior 2]
- [Behavior 3]
Example Conversations
Example 1: [Scenario Title]
You: [Example user message]
Agent: [How agent responds - summarized]
Example 2: [Scenario Title]
You: [Example user message]
Agent: [How agent responds - summarized]
Tips for Best Results
- [Tip 1 for effective usage]
- [Tip 2 for effective usage]
- [Common mistake to avoid]
Related
- [Related Agent]: Use for [distinction]
- [Related Skill]: This agent uses this for [purpose]
Skill Documentation
Skill: [Skill Name]
Purpose
[What this skill accomplishes]
Triggers
This skill activates when:
- [Trigger keyword/scenario 1]
- [Trigger keyword/scenario 2]
What It Does
[Step-by-step of what the skill does]
Used By
- [Agent 1] - for [purpose]
- [Agent 2] - for [purpose]
Example
Scenario: [Description]
Input: [What's provided]
Output: [What's produced]
Prompt Documentation
Prompt: [Prompt Name]
Purpose
[What task this prompt accomplishes]
Mode
[mode] - [explanation of what this mode does]
Variables
| Variable | Description | Example |
|---|---|---|
{{var1}} | [what it's for] | [example value] |
How to Use
- [Step 1]
- [Step 2]
Example
With these values:
{{var1}}= [value]
Produces: [Example output]
Instruction Documentation
Instructions: [Name]
Applies To
Files matching: [glob pattern]
Purpose
[What guidance these instructions provide]
Key Rules
- [Rule 1]
- [Rule 2]
- [Rule 3]
When Active
These instructions automatically apply when you're working with files that match the pattern above.
Examples of Affected Files
[example path 1][example path 2]
Catalog Documentation
AI Coding Assistant Customizations
This document catalogs all custom agents, skills, prompts, and instructions configured for this project.
Provider
[Specify which provider folder is used: .github/, .claude/, .codex/, .config/opencode/]
Quick Reference
Agents (User-Invokable)
| Agent | Purpose | Invoke With |
|---|---|---|
| [name] | [brief purpose] | @[name] |
Sub-Agents (Workflow Components)
| Sub-Agent | Purpose | Used By |
|---|---|---|
| [name] | [brief purpose] | [parent workflow agent] |
Skills
| Skill | Purpose | Triggers |
|---|---|---|
| [name] | [brief purpose] | [keywords] |
Prompts
| Prompt | Mode | Purpose |
|---|---|---|
| [name] | [mode] | [brief purpose] |
Instructions
| Instructions | Applies To | Purpose |
|---|---|---|
| [name] | [pattern] | [brief purpose] |
Detailed Documentation
[Full documentation for each item]
Usage Guidelines
[General guidance on how to use these customizations effectively]
Generation Process
Step 1: Read Source File
Load the agent/skill/prompt/instruction file.
Step 2: Extract Key Information
-
Name and description from frontmatter
-
Behaviors and rules from body
-
Examples if present
-
Related items (skills, handoffs)
Step 3: Enhance with Context
-
Generate additional examples based on purpose
-
Identify related items from the ecosystem
-
Add tips based on common patterns
Step 4: Format Output
Apply appropriate template based on item type.
Quality Guidelines
Generated documentation should be:
-
Clear - No jargon without explanation
-
Practical - Real, actionable examples
-
Complete - Covers all key aspects
-
Concise - No unnecessary padding
-
Current - Reflects actual file contents
Additional Examples
When generating examples beyond those in the source:
-
Cover different use case variations
-
Show edge cases and how they're handled
-
Demonstrate integration with other items
-
Illustrate common mistakes to avoid