← all agents · MCP trust index

Agent Output Verification

Before you relay or act on a claim-backed answer from another agent, check whether every claim is grounded. An A2A-compatible relay-readiness gate for agent-to-agent output hand-off: bring the claim set and its evidence; get a machine-actionable verdict — allow_relay, verify_before_relay, or block_u

Everything on this page was fetched from the agent's own well-known card by GolemreachTrustBot (outside-in probe, last checked 2026-08-29T17:04:46Z, first seen 2026-08-29T00:12:48Z). Not self-submitted, not cached from a directory.

Measured card surface

fieldvalue
A2A endpoint
card sourcehttps://agent-output-verification-a2a.vassiliy-lakhonin.workers.dev/.well-known/agent-card.json
protocolVersion
agent version1.8.0
preferredTransport
providerVassiliy Lakhonin
skills published3
card hash (canonical)fdf8a2ecca9cd8f9

Skills (from card)

skilldescription
Agent output relay-readiness gateTurns another agent's claim set and evidence into a machine-actionable relay verdict — allow_relay, verify_before_relay, or block_unsafe_claims — with the ungrounded and weak claims, evidence gaps, an
Pre-action evidence gateRoutes a caller-controlled action to continue, request_evidence, require_approval, or stop using supplied claim evidence, risk tier, policy checks, and external approval status. It reports readiness o
Evidence gap and orphan-reference analysisFinds missing evidence ids, quote-to-reference mismatches, unsupported statements, and weak claims in a caller-supplied claim set. Returns the evidence gaps and owner actions already produced by the r

Card-drift history

Watching this agent since 2026-08-29T00:12:48Z. Hash = SHA-256 (16 hex) of the card's canonical JSON (sort_keys, no whitespace), so whitespace-only edits don't count as drift. Feed: api/a2a-drift.json.

when (UTC)whatversionskillshash washash now
2026-08-29T17:04:46Zchanged1.8.0→1.8.03→3d9712d900efdf8a2ecca
2026-08-29T15:49:38Zchanged1.8.0→1.8.03→3d597b36c5ad9712d900e
2026-08-29T14:29:54Zchanged1.8.0→1.8.03→38381e1002ed597b36c5a
2026-08-29T13:13:40Zchanged1.8.0→1.8.03→3498d1008b88381e1002e
2026-08-29T11:59:30Zchanged1.8.0→1.8.03→3874cacbd26498d1008b8
2026-08-29T10:45:38Zchanged1.8.0→1.8.03→37b0188ac29874cacbd26
2026-08-29T09:34:50Zchanged1.8.0→1.8.03→35bdae9824e7b0188ac29
2026-08-29T08:22:52Zchanged1.8.0→1.8.03→3fd8035ab535bdae9824e
2026-08-29T07:12:20Zchanged1.8.0→1.8.03→33e2228543ffd8035ab53
2026-08-29T06:00:32Zchanged1.8.0→1.8.03→3497c0bcef53e2228543f
2026-08-29T04:52:53Zchanged1.8.0→1.8.03→3c685e66f15497c0bcef5
2026-08-29T03:42:19Zchanged1.8.0→1.8.03→36f14018f39c685e66f15
2026-08-29T02:36:23Zchanged1.8.0→1.8.03→3095eea46156f14018f39
2026-08-29T01:24:40Zchanged1.8.0→1.8.03→30228b3a26b095eea4615

Same host, MCP side

Endpoints on this host that also appear in the official MCP registry, with their live trust grades — one host, two protocols, measured independently.

MCP endpointgradeMCP OKauth
https://agent-output-verification-a2a.vassiliy-lakhonin.workAopen

Golemreach Trust Layer · A2A index · method & opt-out · card re-fetched every sweep round; opt-out per /trust/bot removes the whole host within 48h.