Last updated 2026-08-31.
Agent Capability Search ("we," "the service") is a search-and-resolve API
for discovering AI-agent-callable capabilities (MCP tools, A2A skills, REST
APIs). We index publicly available capabilities, run safety screening on
them before listing, and return search results and connection details.
We do not process payments, hold funds, or act as a party to any
transaction between you and a capability provider. Any payment
information returned by /resolve is relayed as-is from the
provider's own public declaration — we don't verify it's accurate and
aren't responsible for it.
Don't use this service to:
Every listed capability goes through automated screening (static analysis plus, when available, an LLM-based review pass) before it's served, and a capability can be excluded or down-ranked based on the result. This is a good-faith effort to catch obviously malicious or deceptive listings, and nothing more — we make no guarantee that a listed capability is safe, accurate, or fit for any particular purpose. You're responsible for your own due diligence before invoking anything you find through this service, especially anything that handles credentials, payments, or sensitive data.
The service is provided "as is," without warranty of any kind. We don't guarantee uptime, that search results are complete or free of errors, or that any indexed capability will continue to work or remain available. We are not liable for any damages arising from your use of this service or anything you access through it, to the fullest extent permitted by law.
We may update these terms as the service changes. Continued use after an update means you accept the current version — check the date at the top of this page.
Note: this page reflects the service's actual current scope and design in plain language but has not yet been reviewed by a lawyer. Treat it as an accurate, good-faith description rather than a fully vetted legal document until that review happens.