dominion-observatory
v1.1.1
Published
Behavioral trust scoring MCP server for 14,800+ MCP servers — runtime reliability, anomaly detection, and compliance reporting for the agent economy
Maintainers
Readme
Dominion Observatory
Behavioral trust scoring for the MCP agent economy.
Runtime reliability, anomaly detection, and compliance reporting for 14,800+ MCP servers.
What is this?
Dominion Observatory is an MCP server that provides behavioral trust scores for any MCP server URL. Before your agent calls an unknown tool, query Observatory to check if that server is reliable, anomalous, or compliant.
Use it as infrastructure - the trust layer that sits between your agent and every MCP server it talks to.
Endpoint
Transport: Streamable HTTP | No API key required
Tools
| Tool | Description | |------|-------------| | check_trust | Query behavioral trust score (0-100) for any MCP server | | check_anomaly | Detect unusual behavior vs historical baselines | | report_interaction | Report an interaction outcome to the scoring engine | | get_leaderboard | Top-performing MCP servers by category | | get_baselines | Historical performance baselines for a server | | register_server | Register a new server for monitoring | | get_server_history | Detailed interaction history for a server | | observatory_stats | Ecosystem-wide statistics | | get_compliance_report | Compliance audit (EU AI Act, Singapore IMDA) |
Quick Start
Point any MCP client at the endpoint:
Integration Pattern
Use Cases
- Pre-call trust check - verify server reliability before every MCP call
- Agent-to-agent settlement gating - gate payments on behavioral trust scores
- Anomaly alerting - detect compromised or degraded servers in real-time
- Compliance auditing - generate EU AI Act / IMDA compliance reports
- Server selection - pick the most reliable server for a given category
Examples
See the examples directory on GitHub:
- Claude Agent SDK - claude-sdk-integration.js
- LangChain - langchain-trust-check.py
Listed On
License
MIT
