Product: Agentability
What it does: Public-mode agent readiness evaluator.
Start here: https://agentability.org
API: https://agentability.org/v1
MCP: https://agentability.org/mcp
OpenAPI: https://agentability.org/.well-known/openapi.json
Manifest: https://agentability.org/.well-known/air.json

Verification:
- https://agentability.org/discovery/audit/latest.json

Reports:
- https://agentability.org/reports/{domain}
- https://agentability.org/v1/evaluations/{domain}/latest.json

Spec:
- https://agentability.org/spec.md

How to use the API:
- POST https://agentability.org/v1/evaluate
  {"origin":"https://example.com","profile":"auto"}
- Poll https://agentability.org/v1/runs/{runId} until status is complete
- POST https://agentability.org/mcp (JSON-RPC)

Pillars:
- Discovery: machine entrypoints and canonical docs.
- Callable Surface: usable OpenAPI and callable signals.
- LLM Ingestion: clear, linkable documentation.
- Trust: attestations and provenance signals.
- Reliability: repeatability across fetches.

Failures:
- fail means missing or broken surface.
- warn means partial coverage or unstable responses.

SSRF and abuse protections:
- Only http/https.
- Private IP ranges blocked after DNS resolution.
- Redirects re-checked for private IPs.
- Timeouts and size limits enforced (15s, 2 MB).
- Rate limits apply to public evaluate requests.