MCP Prompt Collector

- 🛠️ Interactive Tools: Provides 4 tools accessible in Claude Desktop for prompt analysis
- 📊 Quality Analysis: Scores prompts on structure, clarity, context, and effectiveness
- 💡 Smart Insights: Identifies patterns and suggests specific improvements
- 📈 Progress Tracking: Shows quality trends and improvement over time
- 🌐 Web Dashboard: Clean interface to view history, analytics, and insights

- ✅ Interactive MCP tools for Claude Desktop integration
- ✅ SQLite database for local storage (no cloud sync)
- ✅ Real-time prompt quality scoring
- ✅ Web dashboard with analytics
- ✅ Automated insights generation

- log_prompt- Log and analyze a prompt with immediate quality feedback
- improve_prompt- Get specific suggestions for improving any prompt
- get_prompt_insights- View analytics and patterns from your prompt history
- view_stats- Display comprehensive statistics and quality trends

- Quality Scoring: 0-10 score based on structure, clarity, and completeness
- Complexity Detection: Categorizes prompts as simple, moderate, or complex
- Technique Suggestions: Recommends optimal prompting techniques
- Issue Detection: Identifies ambiguity, missing context, bias, etc.

- Prompt history with search and filtering
- Quality trends over time
- Complexity distribution charts
- Recent activity overview
- Actionable improvement insights

Once the MCP server is running, you'll have access to these tools in Claude Desktop:

log_prompt - Use: Log and analyze any prompt - Example: "log_prompt" with your prompt text - Returns: Quality score, complexity, and improvement suggestions improve_prompt - Use: Get specific suggestions for a prompt - Example: "improve_prompt" with your prompt and optional goal - Returns: Detailed analysis and actionable improvement recommendations get_prompt_insights - Use: View patterns and insights from your prompt history - Example: "get_prompt_insights" optionally filtered by category - Returns: Analytics on your prompting patterns and common issues view_stats - Use: See comprehensive statistics and trends - Example: "view_stats" for last 30 days (configurable) - Returns: Total prompts, quality trends, complexity breakdown

# Build and run npm run build # Build TypeScript npm run dev # Development mode with auto-reload npm start # Start the MCP server # Code quality npm test # Run tests npm run lint # Check code style npm run format # Format code

- MCP Integration: Registers as an MCP server with Claude Desktop
- Interactive Tools: Provides 4 tools accessible through Claude Desktop interface
- Manual Logging: Users explicitly log prompts using thelog_prompttool
- Analysis: Scores quality, detects patterns, identifies improvements
- Storage: Saves everything locally in SQLite database
- Insights: Generates personalized suggestions based on your patterns
- Dashboard: Provides web interface to view analytics and history

Default configuration is stored in~/.mcp-prompt-collector/config/settings.json:

{ "name": "mcp-prompt-collector", "capabilities": { "intercept": true, "modify": false, "store": true }, "settings": { "enhancePrompts": false, "storageLocation": "~/.mcp-prompt-collector/data", "dashboardPort": 3456, "analysisLevel": "detailed", "retentionDays": 90 } }

All data is stored locally in SQLite database:

- Location:~/.mcp-prompt-collector/data/prompt_history.db
- Privacy: No cloud sync, complete local control
- Retention: Configurable cleanup after N days
- Export: Full data export capabilities

# Clone and setup git clone <repo-url> cd mcp-prompt-collector npm install # Development workflow npm run dev # Watch mode npm run build # Production build npm test # Run tests npm run lint # Code quality # Testing npm run test:unit # Unit tests npm run test:integration # Integration tests

- MCP Protocol Handler: Communicates with Claude Desktop
- Storage Manager: SQLite database operations
- Analysis Engine: Prompt quality scoring and insights
- Dashboard Server: Express.js web interface
- Background Worker: Scheduled analysis and cleanup

- ✅Local Only: All data stays on your machine
- ✅No Cloud Sync: Zero external data transmission
- ✅User Control: Complete data ownership and export
- ✅Transparent: Open source, auditable code

- Node.js: 18+
- Claude Desktop: Latest version with MCP support
- macOS: Primary support (Windows/Linux coming soon)

- Check if port 3456 is available
- Verify Claude Desktop MCP configuration
- Runmcp-prompt-collector statusfor diagnostics

- Make sure you're using thelog_prompttool in Claude Desktop
- Prompts are not automatically captured - you must manually log them
- Check that the MCP server tools are available in Claude Desktop

- Check if server is running on correct port
- Try different port withsetup --port 3457
- Check browser console for errors

MIT License - see LICENSE file for details

This is a web browser that enables your coding agent, such as Claude Code, to visit websites on your behalf and assist you in identifying bugs or creating UI test cases.

MCP server for AI Diagram Maker — generate beautiful software engineering diagrams directly inside Cursor, Claude Desktop, Claude Code, or any MCP-compatible AI agent

Local-first knowledge graph for developers. Watches project files, extracts entities and relationships via LLMs, and lets you query across projects with natural language and source citations.

MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, session continuity, and output truncation to save Claude's context and tokens.

The industry-standard doctrine for safe, observable, and steerable AI agent UX — browse 10 principles, curated examples, and application guides via MCP.

Assists AI developers with requirement clarification, module design, and technical architecture.

Local agent workbench bundling OpenHands, Goose, Aider, and ashlrcode against one local LLM, with ashlr-plugin MCP servers pre-wired.

Connect AI tools like Cursor and VS Code to your product documentation using the Biel.ai platform.

Provides AI assistants with a comprehensive, one-time analysis for complete codebase context and understanding.

A collection of MCP servers that provide cognitive enhancement tools for large language models.