mcp -debug-thinking

This MCP server implements the Debug Thinking framework, which structures each debugging workflow as a persistent knowledge graph. It captures problems, hypotheses, experiments, observations, learnings, and solutions, enabling fast retrieval of past debugging paths and re-use of proven strategies. The server exposes a set of tooling commands through the use_tool interface, including actions to create nodes (problem, hypothesis, experiment, observation, learning, solution) and to query similar problems or recent activity. Typical usage involves defining a debugging problem, formulating hypotheses, running experiments, recording observations, and then extracting learnings to guide future work. You can also search for similar problems and their solutions to speed up debugging paths, or query recent activity to maintain session continuity across runs.

Prerequisites:

• Node.js and npm installed on your system
• Access to a shell/terminal

Installation steps:

1. Install the MCP server globally using npm: npm install -g mcp-server-debug-thinking

2. Verify installation by running the server in development or using the built-in CLI if available. For example, ensure the server exposes the MCP endpoints and can be started with the provided configuration.

3. Create or update your MCP configuration to register the server instance. Example config: { "mcpServers": { "debug-thinking": { "command": "npx", "args": ["mcp-server-debug-thinking"] } } }

4. Start the server using your preferred method (global CLI or local project script) and ensure it can access its storage backend (see repository docs for storage setup).

5. Run tests or sample flows to confirm everything is wired correctly.

Notes:

• If you’re behind a proxy or corporate firewall, ensure npm can access the registry and that the npx invocation can fetch the package as needed.
• Review environment-specific configuration in the project docs if you need to customize storage or logging.

Tips and common issues:

• The server uses a graph-based data model (nodes and edges). When creating a new debugging flow, start by defining the Problem node, then add related Hypothesis, Experiment, and Observation nodes in sequence.
• Use the query action to fetch similar problems and suggested solutions, which can help accelerate debugging paths.
• For environment variables, you may need to configure storage paths or logging levels depending on your deployment. If you encounter storage permission issues, verify the process has read/write access to the configured storage directory.
• Keep your MCP SDK version in sync with the server for best compatibility, as shown by the MCP SDK badge in the README.
• Review the Quick Start guide in the docs/QUICK_START.md for practical, scenario-based usage.