nfragile

The intelligence layer that resolves agent capability invocations

The outside calibration layer for AI agents — verified capability resolution with cryptographically signed trust.

108,035 capabilities across 13,981 AI artifacts. Signed trust passports. Open schema.

Building an agent? Use /api/v1/resolve
Capability-first

Products are decomposed into what they can actually do.

Agent-callable

Agents can query the graph through MCP and API surfaces.

Trust proof

Match Proof and trust passports expose evidence, limits, and risk.

Learning loop

Searches, clicks, feedback, and gaps improve future matches.