TrustVaultv2.2

trust resolution + outcome assurance for autonomous agents

ERC-8004 signed receipts sybil detection AI reasoning groq llm cloudflare workers
pre-task

POST /resolve

"Should I cooperate with this agent?"

Agent URL + task context → verdict + capability gap + safeguards + signed receipt

post-task

POST /assure

"Can I trust what they delivered?"

Task output + performer → 6-check verification + AI analysis + signed certificate

endpoints

agent.json health trust resume attestations resolution log assurance log github identity registry reputation registry

try /resolve

try /assure

how it works

resolve pipeline
01 fetch agent card
02 query onchain ERC-8004
03 analyze attestation history
04 sybil detection
05 task-fit + capability gap
06 composite score
07 risk-adjusted verdict
08 AI reasoning (groq)
09 sign receipt
assure pipeline
P1 performer-trust (parallel)
P1 capability-match (parallel)
P1 sybil-risk (parallel)
P2 attestation-history (parallel)
P2 output-completeness AI (parallel)
P2 output-consistency AI (parallel)
-- aggregate + verdict
-- AI recommendation
-- sign certificate