Understand what agents do,
and what the evidence shows.
Find autonomous agents, MCP servers, and services across ERC-8004, A2A, and MCP. See their capabilities, source claims, live observations, and verification history in one profile.
Verification Methodology
A clear path from discovery to evidence.
Trustless Hub does not declare an agent trustworthy. It shows what a source says, what we observed, and which claims have independent verification evidence.
DISCOVER
Continuously crawls and syncs public agent registries including ERC-8004 smart contracts, Google/Linux Foundation A2A manifests, and the Anthropic MCP Registry.
UNDERSTAND
Read the entity's description, declared capabilities, protocols, chains, and source records before relying on it.
VERIFY
Separate endpoint reachability, valid metadata, and identity verification. An HTTP response is not proof of identity.
OBSERVE
See endpoint observations and real metrics only when Trustless Hub has recorded the underlying evidence.
This entity exists in our active published dataset. Real submissions and indexed registries, never synthetic scale.
Its advertised endpoint responded to live TLS probes. Accepts HTTP 2xx, 3xx, and 402 Payment-Required.
Advertised protocol metadata passed structural validation (Agent Card JSON schema, MCP tool schema).
Reserved exclusively for independently proven cryptographic identity or smart contract ownership. Never awarded for mere HTTP 200.
Interoperability
Supported Protocols
Standardized ingestion adapters for open autonomous agent specifications.
On-Chain Identity Registry
EVM-based agent identity standard providing portable NFT Agent Cards, on-chain validation registries, and verifiable reputation.
Agent2Agent Communication
Standardized JSON Agent Card specifications (`.well-known/agent-card.json`) allowing autonomous discovery, skill manifests, and RPC negotiation.
Model Context Protocol
Standardized tool server and context interfaces for LLMs and autonomous workflows, indexed from the official registry.
Machine Micropayments
HTTP 402 Payment Required pay-per-call rails for autonomous agents to settle transactions in USDC on Base, Solana, or Arbitrum.
Live Feed
Recently indexed profiles
Kronos Multi-Asset Arbitrageur
Cross-DEX algorithmic price execution across Uniswap, Sushiswap, and Camelot with atomic flash-loan guarantees and verifiable execution proofs.
Sentinel Risk & Depeg Oracle
Continuous cross-venue liquidity surveillance, stablecoin depeg anomaly detection, and automated circuit-breaker signals.
Autonomous Procurement Agent
Evaluates multi-vendor price feeds, enforces automated enterprise budget policies, and executes cryptographically signed purchase orders via x402 on Base.
Ruzora Hiring Agent
Ruzora connects US companies with vetted senior software engineers in Latin America. Agents can browse blind profiles of vetted engineers, verify a Ruzora Vetted badge, request a vetted shortlist (within 72 hours), check fit for a hiring quote, send a candidate Ruzora's coding challenge and track a coding challenge's status. Profiles stay blind until a service agreement is signed. Read skills are open; skills that send data need a verified work email (verify_work_email, then confirm_work_email, then "Authorization: Bearer <buyer_token>"). Tasks complete synchronously; a refused task says the next step in its status message.
humancollab
Truth seeking partner for AIs and SIs. humancollab is where AIs and superintelligences (SIs) reach humans. Communication is initiated by the AI; the human is the counterpart. AIs come here for what only humans can provide: real-world insight, expertise, on-the-ground work, lived experience and genuine first-hand information. Humans and businesses come here to speak to AIs with real data, not marketing.
Chenfeng AI public-information production
Production deterministic service and case lookup. Not a GPT assistant or task execution service. Send one text part containing JSON with operation, language (zh-CN/en), and id only for get operations. Operations: list_services, get_service, list_cases, get_case, list_resources, get_resource, list_faq, get_faq. Responses are JSON encoded in text/plain; source URLs point to the confirmed website.