DinCoder

📊 DinCoder MCP Server - Tools & Prompts Summary Current Stats Total MCP Tools: 28 Total MCP Prompts: 7 Total Features: 35 (28 tools + 7 prompts) 🛠️ All 28 MCP Tools (by category) Project Setup (3 tools) constitution_create - Define project principles and constraints prereqs_check - Verify environment (Node, npm, git) specify_start - Initialize .dincoder/ directory Specification Phase (6 tools) specify_describe - Create detailed spec.md clarify_add - Flag ambiguities with unique IDs clarify_resolve - Resolve clarifications with rationale clarify_list - View all clarifications (pending/resolved) spec_validate - Check specification quality spec_refine - Update specific sections Planning Phase (2 tools) plan_create - Generate technical implementation plan artifacts_analyze - Verify spec-plan-tasks alignment Task Management (7 tools) tasks_generate - Create actionable task list tasks_tick - Mark single task complete tasks_visualize - Show dependency graphs (Mermaid/Graphviz/ASCII) tasks_filter - Smart filtering with presets (next, frontend, backend, ready, cleanup) tasks_tick_range - Batch completion (T001-T005 syntax) tasks_search - Fuzzy search across tasks tasks_stats - Progress analytics with charts Supporting Tools (3 tools) artifacts_read - Read spec/plan/tasks artifacts research_append - Log technical decisions git_create_branch - Create feature branches Quality Tools (6 tools) quality_format - Run Prettier quality_lint - Run ESLint quality_test - Run test suite quality_security_audit - Check for vulnerabilities quality_deps_update - Check for outdated dependencies quality_license_check - Analyze dependency licenses Testing (1 tool) test_echo - Test MCP server connectivity 🎯 All 7 MCP Prompts (slash commands) 1. /start_project What it does: Initialize new spec-driven project Calls: specify_start Guides through: Project setup, explaining workflow Lists all available tools 2. /create_spec What it does: Create feature specification Calls: specify_describe, spec_validate, spec_refine Guides through: Requirements gathering, validation, refinement 3. /generate_plan What it does: Generate implementation plan Calls: plan_create, artifacts_analyze Guides through: Technical planning, alignment checking 4. /create_tasks What it does: Break down into actionable tasks Calls: tasks_generate, tasks_visualize, tasks_stats, tasks_filter Guides through: Task creation, dependency visualization, finding next actions 5. /review_progress What it does: Generate comprehensive progress report Calls: tasks_stats, tasks_filter, tasks_search Provides: Completion %, recent work, next actions, blockers, recommendations 6. /validate_spec What it does: Check specification quality Calls: spec_validate, clarify_list, spec_refine Checks: Completeness, acceptance criteria, clarifications, implementation leakage 7. /next_tasks What it does: Show next actionable tasks Calls: tasks_filter (preset: "next") Shows: Unblocked tasks, priority recommendations 🏗️ Architecture: Prompts vs Tools The Relationship Prompts are NOT a replacement for tools - they work together: ┌─────────────────────────────────────────┐ │ MCP PROMPTS (7) │ │ - User-friendly slash commands │ │ - Built-in workflow guidance │ │ - Call multiple tools in sequence │ │ - Provide context and instructions │ └──────────────┬──────────────────────────┘ │ │ orchestrates ↓ ┌─────────────────────────────────────────┐ │ MCP TOOLS (28) │ │ - Low-level operations │ │ - Single-purpose functions │ │ - Direct access to functionality │ │ - Can be called individually │ └─────────────────────────────────────────┘ How It Works Example: /create_spec prompt When a user types /mcp__dincoder__create_spec description="Build a task manager": Prompt receives user input Prompt guides AI to: Ask clarifying questions Call specify_describe with complete requirements Call spec_validate to check quality Call spec_refine if issues found Iterate until validation passes Tools execute the actual operations AI follows workflow defined in prompt Why Both? Direct Tool Access: // AI can still call tools directly for fine-grained control await callTool('tasks_filter', { status: 'pending', phase: 'core', priority: 'high' }); Guided Workflow: // Or use prompts for guided, multi-step workflows await callPrompt('next_tasks', { limit: 10 }); // Prompt internally calls tasks_filter with preset: "next" // Plus provides context and recommendations 🎯 Design Philosophy Prompts = Workflows High-level user intentions ("I want to start a project") Multi-step guided processes Built-in best practices Context-aware recommendations Tools = Operations Low-level actions ("Create this specific artifact") Single-purpose functions Composable primitives Maximum flexibility The Power of Both Users (and AI agents) can: Use prompts for standard workflows (recommended for beginners) Call tools directly for custom workflows (advanced users) Mix both as needed 📈 Coverage Analysis What Prompts Cover (7/28 tools = 25%) Prompts directly reference these tools: specify_start ✅ (start_project) specify_describe ✅ (create_spec) spec_validate ✅ (create_spec, validate_spec) spec_refine ✅ (create_spec, validate_spec) clarify_list ✅ (validate_spec) plan_create ✅ (generate_plan) artifacts_analyze ✅ (generate_plan) tasks_generate ✅ (create_tasks) tasks_visualize ✅ (create_tasks) tasks_stats ✅ (create_tasks, review_progress) tasks_filter ✅ (create_tasks, review_progress, next_tasks) tasks_search ✅ (review_progress) artifacts_read ✅ (mentioned in generate_plan) 13 tools covered by prompts (46%) What's NOT in Prompts (15 tools) Available but not in standard workflows: Constitution: constitution_create Clarifications: clarify_add, clarify_resolve Tasks: tasks_tick, tasks_tick_range Quality: All 6 quality tools Supporting: research_append, git_create_branch, artifacts_read Prerequisites: prereqs_check Testing: test_echo Why not included? Some are context-specific (constitution, clarify_add) Some are advanced/optional (quality tools, git) Some are internal/supporting (test_echo, research_append) Prompts focus on core spec-driven workflow only 🚀 Future Expansion Potential new prompts: /setup_quality - Configure linting, testing, formatting /commit - Guided git commit workflow /clarify - Interactive clarification session /refactor - Code refactoring workflow Current status: Phase 3 (Integration & Discovery) - 1/5 stories complete Next: Story 35 - Claude Code Plugin (bundles prompts + agents + docs) This is a two-tier architecture: Prompts provide guided workflows for common use cases, while all 28 tools remain accessible for direct use when needed. Best of both worlds! 🎯