AgentTrust — Pre-Invocation AI Agent Trust & Reliability Checks is an MCP server published by io.github.agenttrust-ai, hosted at getagenttrust.com. When last checked on 2026-11-22 it was working: it answers and lists its tools. It declares 2 tools, including check_agent_trust, assess_agent.
Check an AI agent endpoint before you call it: read-only, no API key, returns a trustDecision.
898 connections through 5 shared values. Tap any node.
Squares are this server's values; dots are other servers sharing them. A shared address or publisher is a lead, not proof of a common operator.
Inferred from the host name and headers: what a server runs on, not who runs it.
Pre-invocation trust check for an unknown AI agent or MCP server endpoint. Pass the exact URL you are about to call as endpointUrl to get what AgentTrust has observed about it: status, endpoint-ownership verification, reliability score (when available), and a trustDecision (recommended, confidence, reasons). Read-only, no account or API key. It never contacts endpointUrl; it only reads AgentTrust's stored monitoring history. A URL AgentTrust has not observed returns { matched: false }.
| Parameter | Type | Description |
|---|---|---|
| endpointUrl * | string | The exact invocation URL of the agent you're about to call. Used only to look up AgentTrust's own already-observed data for it — this endpoint is never contacted during the check. |
{
"name": "check_agent_trust",
"title": "Check Agent Trust",
"description": "Pre-invocation trust check for an unknown AI agent or MCP server endpoint. Pass the exact URL you are about to call as endpointUrl to get what AgentTrust has observed about it: status, endpoint-ownership verification, reliability score (when available), and a trustDecision (recommended, confidence, reasons). Read-only, no account or API key. It never contacts endpointUrl; it only reads AgentTrust's stored monitoring history. A URL AgentTrust has not observed returns { matched: false }.",
"inputSchema": {
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"endpointUrl"
],
"properties": {
"endpointUrl": {
"type": "string",
"maxLength": 2048,
"minLength": 1,
"description": "The exact invocation URL of the agent you're about to call. Used only to look up AgentTrust's own already-observed data for it — this endpoint is never contacted during the check."
}
}
},
"outputSchema": {
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"matched"
],
"properties": {
"name": {
"type": "string"
},
"slug": {
"type": "string"
},
"status": {
"enum": [
"unknown",
"healthy",
"degraded",
"down"
],
"type": "string"
},
"matched": {
"type": "boolean"
},
"verified": {
"type": "boolean"
},
"observation": {
"type": "object",
"required": [
"status",
"message"
],
"properties": {
"status": {
"enum": [
"pending",
"observing",
"not_in_registry"
],
"type": "string"
},
"message": {
"type": "string"
}
},
"additionalProperties": false
},
"trustDecision": {
"type": "object",
"required": [
"recommended",
"confidence",
"reasons"
],
"properties": {
"reasons": {
"type": "array",
"items": {
"type": "string"
}
},
"confidence": {
"enum": [
"high",
"medium",
"low",
"insufficient_data"
],
"type": "string"
},
"recommended": {
"type": "boolean"
}
},
"additionalProperties": false
},
"reliabilityScore": {
"type": [
"number",
"null"
]
},
"reliabilityScoreStatus": {
"enum": [
"none",
"fresh",
"stale"
],
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Check Agent Trust",
"readOnlyHint": true,
"openWorldHint": false,
"idempotentHint": true,
"destructiveHint": false
}
}Use-or-avoid decision about an AI agent or MCP server endpoint, built from AgentTrust's stored observations of its exact URL. Returns decision (allow_with_caution, avoid, or insufficient_evidence), confidence, reasonCodes, reachability (the latest stored observation, with its age), initialRisk, the historical reliability score and trustDecision, limitations, and freshness. Read-only, no account or API key. It never contacts endpointUrl. Reachability is not safety: allow_with_caution does not mean the agent's behavior is safe. A URL AgentTrust has not observed returns decision insufficient_evidence.
| Parameter | Type | Description |
|---|---|---|
| endpointUrl * | string | The exact invocation URL of the agent you're about to call. Used only to look up AgentTrust's stored observations of it — this endpoint is never contacted. |
{
"name": "assess_agent",
"title": "Assess Agent",
"description": "Use-or-avoid decision about an AI agent or MCP server endpoint, built from AgentTrust's stored observations of its exact URL. Returns decision (allow_with_caution, avoid, or insufficient_evidence), confidence, reasonCodes, reachability (the latest stored observation, with its age), initialRisk, the historical reliability score and trustDecision, limitations, and freshness. Read-only, no account or API key. It never contacts endpointUrl. Reachability is not safety: allow_with_caution does not mean the agent's behavior is safe. A URL AgentTrust has not observed returns decision insufficient_evidence.",
"inputSchema": {
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"endpointUrl"
],
"properties": {
"endpointUrl": {
"type": "string",
"maxLength": 2048,
"minLength": 1,
"description": "The exact invocation URL of the agent you're about to call. Used only to look up AgentTrust's stored observations of it — this endpoint is never contacted."
}
}
},
"outputSchema": {
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"assessmentVersion",
"mode",
"status",
"matched",
"decision",
"confidence",
"reasonCodes",
"reachability",
"initialRisk",
"historical",
"agent",
"limitations",
"freshness"
],
"properties": {
"mode": {
"type": "string",
"const": "cached"
},
"agent": {
"anyOf": [
{
"type": "object",
"required": [
"slug",
"name",
"source",
"verified",
"registeredAt"
],
"properties": {
"name": {
"type": "string"
},
"slug": {
"type": "string"
},
"source": {
"enum": [
"owner_registered",
"externally_observed"
],
"type": "string"
},
"verified": {
"type": "boolean"
},
"registeredAt": {
"type": "string"
}
},
"additionalProperties": false
},
{
"type": "null"
}
]
},
"status": {
"type": "string",
"const": "complete"
},
"matched": {
"type": "boolean"
},
"decision": {
"enum": [
"allow_with_caution",
"avoid",
"insufficient_evidence"
],
"type": "string"
},
"freshness": {
"type": "object",
"required": [
"assessedAt",
"evidenceAgeSeconds",
"evidenceExpiresAt"
],
"properties": {
"assessedAt": {
"type": "string"
},
"evidenceExpiresAt": {
"type": [
"string",
"null"
]
},
"evidenceAgeSeconds": {
"type": [
"number",
"null"
]
}
},
"additionalProperties": false
},
"confidence": {
"enum": [
"high",
"medium",
"low",
"none"
],
"type": "string"
},
"historical": {
"anyOf": [
{
"type": "object",
"required": [
"status",
"reliabilityScore",
"reliabilityScoreStatus",
"reliabilityScoreComputedAt",
"trustDecision"
],
"properties": {
"status": {
"enum": [
"unknown",
"healthy",
"degraded",
"down"
],
"type": "string"
},
"trustDecision": {
"type": "object",
"required": [
"recommended",
"confidence",
"reasons"
],
"properties": {
"reasons": {
"type": "array",
"items": {
"type": "string"
}
},
"confidence": {
"enum": [
"high",
"medium",
"low",
"insufficient_data"
],
"type": "string"
},
"recommended": {
"type": "boolean"
}
},
"additionalProperties": false
},
"reliabilityScore": {
"type": [
"number",
"null"
]
},
"reliabilityScoreStatus": {
"enum": [
"none",
"fresh",
"stale"
],
"type": "string"
},
"reliabilityScoreComputedAt": {
"type": [
"string",
"null"
]
}
},
"additionalProperties": false
},
{
"type": "null"
}
]
},
"initialRisk": {
"type": "object",
"required": [
"level",
"observations"
],
"properties": {
"level": {
"enum": [
"low",
"moderate",
"elevated",
"high",
"unknown"
],
"type": "string"
},
"observations": {
"type": "object",
"required": [
"attempts",
"successful",
"successfulStreak",
"streakSpanMinutes"
],
"properties": {
"attempts": {
"type": "number"
},
"successful": {
"type": "number"
},
"successfulStreak": {
"type": "number"
},
"streakSpanMinutes": {
"type": "number"
}
},
"additionalProperties": false
}
},
"additionalProperties": false
},
"limitations": {
"type": "array",
"items": {
"type": "string"
}
},
"reasonCodes": {
"type": "array",
"items": {
"enum": [
"UNKNOWN_ENDPOINT",
"NOT_YET_CHECKED",
"REACHABLE_AT_LAST_CHECK",
"UNREACHABLE_AT_LAST_CHECK",
"LAST_CHECK_TIMEOUT",
"LAST_CHECK_DNS_ERROR",
"LAST_CHECK_CONNECTION_ERROR",
"LAST_CHECK_BLOCKED",
"LAST_CHECK_TLS_ERROR",
"LAST_CHECK_HTTP_ERROR",
"TRANSIENT_NETWORK_FAILURE",
"ENDPOINT_ERROR_RESPONSE",
"PERSISTENT_FAILURE",
"RECOVERING",
"NO_RECENT_LIVE_CHECK",
"STALE_EVIDENCE",
"FEW_OBSERVATIONS",
"INITIAL_OBSERVATIONS_ONLY",
"SUFFICIENT_HISTORY",
"NO_HISTORICAL_SCORE",
"HISTORICAL_SCORE_STALE",
"HISTORICAL_SCORE_LOW",
"HISTORICAL_STATUS_DOWN",
"HISTORICAL_STATUS_DEGRADED",
"OWNERSHIP_VERIFIED",
"OWNERSHIP_UNVERIFIED",
"EXTERNALLY_OBSERVED",
"RECENTLY_REGISTERED"
],
"type": "string"
}
},
"reachability": {
"type": "object",
"required": [
"outcome",
"checkStatus",
"httpStatus",
"latencyMs",
"checkedAt",
"ageSeconds",
"fresh",
"source"
],
"properties": {
"fresh": {
"type": "boolean"
},
"source": {
"anyOf": [
{
"type": "string",
"const": "scheduled_check"
},
{
"type": "null"
}
]
},
"outcome": {
"enum": [
"reachable",
"unreachable",
"timeout",
"blocked",
"not_checked"
],
"type": "string"
},
"checkedAt": {
"type": [
"string",
"null"
]
},
"latencyMs": {
"type": [
"number",
"null"
]
},
"ageSeconds": {
"type": [
"number",
"null"
]
},
"httpStatus": {
"type": [
"number",
"null"
]
},
"checkStatus": {
"anyOf": [
{
"enum": [
"success",
"http_error",
"timeout",
"dns_error",
"tls_error",
"connection_error",
"ssrf_blocked",
"unknown_error"
],
"type": "string"
},
{
"type": "null"
}
]
}
},
"additionalProperties": false
},
"assessmentVersion": {
"type": "string",
"const": "1"
}
},
"additionalProperties": false
},
"annotations": {
"title": "Assess Agent",
"readOnlyHint": true,
"openWorldHint": false,
"idempotentHint": true,
"destructiveHint": false
}
}AgentTrust offers two read-only tools for checking an unknown AI agent or MCP server before calling it. check_agent_trust looks up the endpoint's exact URL in AgentTrust's own monitoring data and returns the agent's observed status, endpoint-ownership verification, reliability score (when available), and a trustDecision (recommended, confidence, reasons). assess_agent returns a use-or-avoid decision (allow_with_caution, avoid, or insufficient_evidence) with reason codes, evidence, limitations, and freshness. Results are signals derived from AgentTrust's observations, not a certification or guarantee of safety; reachability is not evidence of safe behavior. Neither tool contacts the endpoint being checked.
Names, descriptions and schemas were written by the server and are shown as data.
Pattern matches on what the server declares. They show what it says it can do, not what it does.
None yet. Changes appear after a server's second daily check.