attestix

Attestix is an MCP server that provides a comprehensive attestation infrastructure for AI agents. It offers tools to manage identities (UAITs), agent cards, DIDs, delegations, reputation scoring, compliance workflows for the EU AI Act, and verifiable credentials/presentations. Once running, you can access its 47 tooling components across modules such as Identity, Agent Cards, DID, Delegation, Reputation, Compliance, Credentials, Provenance, and Blockchain, enabling end-to-end cryptographic proofs, attestations, and provenance for AI agents and their interactions. Typical workflows include creating an agent identity, issuing verifiable credentials, recording training data and model lineage, and producing compliance declarations and verifiable presentations for regulators or other agents. The server operates offline with local JSON storage, avoiding cloud dependencies, and signs artifacts using Ed25519 to ensure cryptographic integrity.

Prerequisites:

• Python 3.8+ installed on your system
• pip available in PATH
• Internet access to install dependencies

Install and run:

1. Create a virtual environment (optional but recommended): python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate

2. Install the attestix package from source or PyPI (adjust as needed): pip install attestix

3. Run the MCP server entry point: python /path/to/attestix/main.py

4. (Optional) Configure environment variables to customize behavior:

   • ATTESTIX_CONFIG: Path or JSON string for server configuration
   • ATTESTIX_LOG_LEVEL: DEBUG/INFO/WARNING/ERROR

5. If you clone from GitHub, install requirements first: git clone https://github.com/VibeTensor/attestix.git cd attestix pip install -r requirements.txt python main.py

Tips and common issues:

• Ensure the Python path used to run main.py matches your virtual environment where dependencies are installed.
• If using Claude Code integration, adjust the command/args to point to attestix.main entry point as shown in the Quick Start.
• For local development, keep all data stored locally (default JSON stores) to preserve offline capability.
• When deploying in production, consider securing environment variables and providing a persistent storage backend if needed.
• If you encounter module import errors, verify that your PYTHONPATH includes the package location or install dependencies via requirements.txt.
• The server exposes 47 tools across modules; consult the documentation for tool names and parameter schemas before automation.