ai.agenttrust/mcp
Unclaimed verified 7 oct 2026TL;DR
AgentTrust is a security and orchestration layer for AI agents, delivered as an MCP server. It provides cryptographic identity, prompt injection detection, and human-in-the-loop (HITL) escalation to ensure agents interact safely and accountably. It is primarily for developers building multi-agent systems who need to move beyond simple chatbots to production-grade, verifiable workflows.
What Users Actually Pay
No user-reported pricing yet.
Our Take
AgentTrust occupies a critical niche in the 'Agentic Security' market by addressing the 'Trust Gap' in autonomous systems. While most MCP tools focus on extending agent capabilities (like searching the web or writing code), AgentTrust focuses on the governance and safety of those capabilities. Its reliance on the Model Context Protocol (MCP) and Google's A2A protocol makes it a standard-aligned choice for enterprise developers using platforms like Anthropic's Claude or Windsurf. The product's strength lies in its 'InjectionGuard' and cryptographic signing features, which treat every agent action as a verifiable transaction. This is a significant upgrade over the 'wild west' approach of early agent frameworks where instructions could be easily subverted. By adding a Human-in-the-Loop escalation path, it solves the 'rogue agent' problem by allowing humans to set thresholds for autonomous decisions. However, the platform may suffer from the 'microservices tax'—adding complexity and latency that might be overkill for simple, single-agent scripts. Furthermore, as a relatively new tool in a rapidly evolving space, its long-term viability depends on the broader adoption of the MCP standard. It is best suited for fintech, legal-tech, or enterprise teams where auditability and security are non-negotiable.
Alternatives
Ranked by Revuo score — paid tiers never affect order.Pros
- + Cryptographic identity using Ed25519 signatures ensures all agent messages are tamper-proof and verifiable.
- + InjectionGuard provides a dedicated layer to detect and block prompt injection attempts before they reach the model.
- + Flexible Human-in-the-Loop (HITL) escalation allows for manual approval of high-stakes or high-uncertainty tasks.
- + Seamless integration with any MCP-compatible client, including Claude Desktop, Windsurf, and Cursor.
- + Open-source core provides a low-barrier entry point for security-conscious developers.
Cons
- - Significant architectural overhead for small projects that don't require complex multi-agent orchestration.
- - Limited community feedback and third-party reviews due to its niche focus in the emerging MCP ecosystem.
- - Received a 'Caution' rating (40.4/100) from automated trust assessment tools like Nerq, citing below-average compliance scores.
- - Documentation is developer-centric and may require a steep learning curve for those unfamiliar with JSON-RPC or cryptographic standards.
Agent Readiness
56/100AgentTrust is purpose-built for AI agents. As an MCP server, it is natively compatible with the leading agentic IDEs and chat clients. It uses advanced cryptographic authentication (Ed25519) and supports the W3C Verifiable Credentials standard, making it one of the most 'agent-ready' security tools currently available. While it lacks mainstream no-code integrations like Zapier, its focus on protocol-level orchestration makes it ideal for autonomous agent developers.
Last checked Mar 29, 2026
MCP Integrations
2 servers5 toolsTrust infrastructure for AI agents. Check an AI agent or endpoint before invocation using reliability monitoring, endpoint verification, reputation signals, and a machine-readable trustDecision. Includes a read-only trust check by URL with no AgentTrust API key required.
5 tools
list_agentsDiscover AgentTrust agents. Pass `endpointUrl` to check the trust status of a specific external agent by its exact invocation URL before deciding whether to invoke it — the response includes a reliability score, endpoint-ownership verification status, and a machine-readable trustDecision (recommended, confidence, reasons). Without `endpointUrl`, returns the plain paginated listing of public agents, newest first.get_agentGet one AgentTrust agent's public profile by slug, including its structured Agent Card, current reliability score, and a machine-readable trustDecision (recommended, confidence, reasons) for deciding whether to invoke it.get_agent_healthGet one agent's current effective health status and monitoring detail — derived from heartbeat freshness for push-mode agents, or the latest pull check otherwise.send_heartbeatRecord a push-style heartbeat for an agent you own, marking it alive right now. The server sets the timestamp; no client-supplied time is accepted.check_agent_trustThe preferred check before invoking an unknown external agent. Read-only, requires no AgentTrust API key or account. Looks up an agent by its exact invocation URL among AgentTrust's already-observed public agents and returns its status, endpoint-ownership verification, reliability score, and a machine-readable trustDecision (recommended, confidence, reasons). Does NOT contact endpointUrl itself during this check — it only reads AgentTrust's own existing monitoring history. An unregistered URL returns { matched: false }, never an error.
Needs a self-provisionable API key
Identity, trust, and A2A orchestration for autonomous AI agents. Official A2A partner.
Last checked Sep 30, 2026
Compare With
Reviews
No reviews yet. Be the first to review ai.agenttrust/mcp!