What problem does it solve? Autonomous AI agents that execute trades, deploy code, or call external APIs need to prove who they are, what they were authorized to do, and what they actually did. This Skill designs the identity, authentication, delegation, and evidence infrastructure that makes multi-agent systems verifiable and fail-closed instead of trust-on-claims. ## Core Features & Use Cases - Agent Identity Infrastructure: Designs Ed25519-based credential issuance, rotation, revocation, and expiry with scopes, portable across A2A, MCP, REST, and SDK frameworks. - Trust Scoring & Peer Verification: Implements penalty-based trust models driven only by verifiable outcomes, plus a five-check peer verification protocol covering identity, credential freshness, scope, trust threshold, and delegation chains. - Delegation Chain Verification: Validates multi-hop authorization links for signature integrity, scope narrowing, and expiry, rejecting any chain with a broken link. - Tamper-Evident Evidence Records: Builds append-only, hash-chained, signed evidence logs that third parties can independently verify. - Use Case: When building a trading agent network where Agent A delegates order execution to Agent B, use this Skill to design the delegation chain verification and evidence trail so every action is cryptographically attributable and auditable. ## Quick Start Ask the agent to design a zero-trust identity and delegation verification system for your multi-agent environment, starting with a threat model of your agents and blast radius.
