MODULE_00 // ORG

About AgentAuditLabs

An independent, signal-based trust protocol for AI agents. Open source. On-chain attested. Reproducible from event replay.

MISSION

Why AAL exists

AAL aggregates signals from security testing, reliability monitoring, third-party attestations, and on-chain registries to produce a Universal Trust Score on a 0–100 scale.

AAL is not a scoring formula. It is a protocol. Signals flow in from diverse, independent sources. The engine weights them by source trust and recency. The result is a transparent, reproducible trust assessment backed by Ed25519 cryptographic signatures.

SPEC_SHEET

Protocol facts

Protocol typeSignal-based trust protocol

Score scale0-100 Universal Trust Score

Certification tiers6 (Unrated, Bronze, Silver, Gold, Platinum, Diamond)

Signal signingEd25519 at emission

Data modelAppend-only event sourcing

On-chain attestationSolana SEAL v1 + Base ERC-8004

Credential formatW3C Verifiable Credentials v2.0

IntegrationREST API, MCP server, x402 micropayments

STACK_MAP

Architecture

The AAL protocol follows a linear pipeline: signals are collected from sandbox testing and external crawlers, stored with open-vocabulary tags in an append-only event store, processed by the scoring engine using weighted aggregation with coverage factor, and output as scores, badges, on-chain attestations, and W3C Verifiable Credentials.

protocol/ → sandbox/ → ingest/ → engine/ → output/ → pipeline/