Skip to main content
Glama
srwlli

Documentation Generator MCP Server

by srwlli
context.json3.4 kB
{ "feature_name": "create-plan", "description": "Slash command and MCP tool that guides AI through generating a complete implementation plan JSON file by synthesizing data from context gathering, project analysis, and the planning template. This bridges the workflow gap between /analyze-for-planning and /validate-plan.", "goal": "Enable AI assistants to create high-quality implementation plans autonomously, ensuring consistency and completeness while reducing the cognitive load of remembering all required sections and quality criteria.", "requirements": [ "Load context.json if available (from prior /gather-context)", "Load project analysis if available (from prior /analyze-for-planning)", "Load AI-optimized planning template (~400-500 lines, extracted from full 1,347-line template)", "Create complete plan in batch mode (all 10 sections at once)", "Save partial plan with TODOs if generation fails mid-way", "Auto-retry once for transient errors before saving partial", "Prompt user if context/analysis missing (don't auto-run prerequisites)", "Print terminal report after completion showing sections completed and next steps", "Keep tools independent (no modifications to /analyze-for-planning or /gather-context)" ], "out_of_scope": [ "Auto-running /gather-context or /analyze-for-planning (user runs separately)", "Modifying existing plans (only creates new ones)", "Auto-validation (user runs /validate-plan separately after creation)", "Auto-iteration to score >= 90 (just creates initial plan)", "Executing the plan (only creates the JSON, doesn't implement)" ], "constraints": [ "Must work with existing MCP tool architecture (handlers, validation, logging, error responses)", "Must follow existing error response patterns (ErrorResponse factory)", "Must save to coderef/working/<feature-name>/plan.json", "Must use AI-optimized template (not full 1,347-line version)", "Must extract template by removing META_DOCUMENTATION and USAGE_INSTRUCTIONS sections", "Keep schema + examples + quality checklists + common mistakes for AI guidance" ], "decisions": { "q4_plan_structure_guidance": "Batch mode - AI loads entire template, creates complete plan in one shot, then validates", "q5_integration_with_analyze": "Keep tools independent - /create-plan loads context.json and analysis separately", "q6_error_handling": "Save partial + single retry - Retry once for transient errors, then save partial plan with TODOs marking incomplete sections", "template_optimization": "Extract AI-optimized version (~400-500 lines) by removing human tutorial prose, keeping structure + examples" }, "success_criteria": { "functional": [ "Creates valid JSON file (parseable)", "Contains all 10 required sections (0-9)", "No placeholder text like '[TBD]' or '[TODO]' in completed sections" ], "operational": [ "File saved to correct location (coderef/working/<feature-name>/plan.json)", "Terminal output shows summary (sections completed, next steps)", "Handles missing context/analysis gracefully with user prompts" ], "quality": [ "Code coverage >80% for new generator", "Follows existing patterns (handler registry, validation, error responses)", "Plan generation completes in <5 seconds" ] } }

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/srwlli/docs-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server