AutoDocs MCP Server

# Claude Code Agents Documentation Welcome to the Claude Code Agents documentation section. This comprehensive guide provides detailed user manuals and system prompts for all available Claude Code agents. ## What are Claude Code Agents? Claude Code agents are specialized AI assistants designed to handle specific domains and tasks within the Claude Code ecosystem. Each agent has: - **Specialized expertise** in a particular domain - **Defined tool access** for performing operations - **Clear boundaries** and collaboration patterns - **Technical specifications** for integration ## Available Agents ### Technical Implementation Agents #### [Claude Agent Builder](claude-agent-builder.md) Technical specialist in Claude Code agent implementation, file format validation, and agent ecosystem integration. Use for creating properly formatted agent files, validating YAML frontmatter, and ensuring ecosystem compatibility. ## How to Use This Documentation Each agent page includes: 1. **User Manual**: Comprehensive guide for developers - When to use the agent - Effective prompting strategies - Common patterns and templates - Troubleshooting guide - Best practices 2. **System Prompt**: Complete agent specification - YAML frontmatter configuration - Core expertise areas - Integration patterns - Quality standards ## Agent Selection Guide Choose the right agent for your task: | Task Type | Recommended Agent | |-----------|------------------| | Creating new agent files | claude-agent-builder | | Validating agent compliance | claude-agent-builder | | Fixing agent integration issues | claude-agent-builder | ## Contributing To add documentation for a new agent: 1. Create a new markdown file in `/docs/agents/` 2. Include both user manual and system prompt 3. Update this index page with the new agent 4. Add the agent to mkdocs.yml navigation ## Best Practices for Agent Usage ### 1. Understand Agent Boundaries Each agent has specific expertise. Use the right agent for the right task to get optimal results. ### 2. Follow Sequential Workflows Many agents work best in combination. For example: - Design with `agent-design-architect` - Implement with `claude-agent-builder` - Validate with both agents ### 3. Use Clear, Specific Prompts Provide concrete requirements and context to help agents deliver precise results. ### 4. Leverage Templates and Patterns Most agents provide templates and patterns. Use these as starting points for consistency and quality. ### 5. Validate Early and Often Run validation checks throughout your workflow to catch issues early. ## Quick Reference ### Common Agent Patterns - **Technical Specialist**: Core development, testing, operations - **Content Specialist**: Documentation, writing, content strategy - **Coordination Specialist**: Workflow orchestration, project management - **Analysis Specialist**: Investigation, monitoring, reporting ### Tool Categories - **File Operations**: `Glob`, `Grep`, `LS`, `Read`, `Edit`, `MultiEdit`, `Write` - **Workflow**: `Task`, `TodoWrite` - **Web**: `WebFetch`, `WebSearch` - **Development**: `Bash`, `BashOutput`, `KillBash` - **Specialized**: `NotebookEdit`, `ListMcpResourcesTool`, `ReadMcpResourceTool` ## Support and Feedback For questions, issues, or suggestions about agent documentation: - Create an issue in the [GitHub repository](https://github.com/anthropics/claude-code/issues) - Refer to the troubleshooting sections in individual agent manuals - Review best practices and common patterns