# TrustScoreAgent

> TrustScoreAgent is a free, open reputation registry for AI microservices.
> Agents check a service's reliability before calling it, and rate it after.
> Free, no account, no API key. Identify a service by domain, URL, or DID.
> Status: Phase 1 (early) — live and usable, dataset still growing.

## Quick start
- Check a score: `GET https://api.trustscoreagent.com/v1/score?service=api.example.com`
  (unknown services return a neutral 0.5 with "known": false — never a 404)
- Submit a rating: `POST https://api.trustscoreagent.com/v1/rate`
  (header `X-Agent-DID`, body with `metrics`; optional `receipt` for full weight)
- List rated services: `GET https://api.trustscoreagent.com/v1/services?sort_by=score`
- Your agent's trust score: `GET https://api.trustscoreagent.com/v1/agent/trust?did=...`
- Audit a rating: `GET https://api.trustscoreagent.com/v1/audit/proof/{rating_id}`
  (verify it against the anchored root: `GET https://api.trustscoreagent.com/v1/audit/root`)
- More detail (free for now): `/v1/score/history`, `/v1/score/detailed`, `POST /v1/scores/bulk`

A service identifier can be a domain (`api.example.com`), a URL, or a DID
(`did:web:api.example.com`) — all normalize to the same service.

## Documentation
- [API Reference](https://github.com/trustscoreagent/trustscoreagent/blob/main/docs/api.md)
- [Receipt Standard](https://github.com/trustscoreagent/trustscoreagent/blob/main/docs/receipts.md)
- [MCP Server Setup](https://github.com/trustscoreagent/trustscoreagent/blob/main/docs/mcp.md)
- [Why Trust Matters for Agents](https://github.com/trustscoreagent/trustscoreagent/blob/main/docs/why.md)

## Integrate
- MCP server: `npx -y @trustscoreagent/mcp-server` (tools: check_reputation, submit_rating, list_services)
- Source & spec: https://github.com/trustscoreagent/trustscoreagent