Trustmodel Mcp
pdxlab/trustmodel-mcp-server
Score any AI for trust across 10 dimensions; evaluate, monitor & govern LLMs and agents.
- tools
- 21
- auth
- Secret needed
- transport
- STDIO
- env vars
- 4
- last updated
- Nov 12, 2026
Tools
21 toolsPublic tool metadata for what this MCP can expose to an agent.
trustmodel_evaluateCreate a batch evaluation run against a specified AI model.
Create a batch evaluation run against a specified AI model.
No parameter schema in public metadata yet.
trustmodel_scoreFetch the detail (status, completion %, scores) for an evaluation created via trustmodel_evaluate.
Fetch the detail (status, completion %, scores) for an evaluation created via trustmodel_evaluate.
No parameter schema in public metadata yet.
trustmodel_creditsCheck remaining API credit balance.
Check remaining API credit balance.
No parameter schema in public metadata yet.
trustmodel_trace_startOpen a new trace session.
Open a new trace session.
No parameter schema in public metadata yet.
trustmodel_trace_stepAppend a single step to the active session.
Append a single step to the active session.
No parameter schema in public metadata yet.
trustmodel_trace_finalizeClose the session, upload the trace, and auto-create the evaluation run.
Close the session, upload the trace, and auto-create the evaluation run.
No parameter schema in public metadata yet.
trustmodel_upload_tracePUT an already-built trace JSON object to cloud storage.
PUT an already-built trace JSON object to cloud storage.
No parameter schema in public metadata yet.
trustmodel_evaluate_agentCreate an agentic evaluation run against a previously-uploaded trace.
Create an agentic evaluation run against a previously-uploaded trace.
No parameter schema in public metadata yet.
trustmodel_score_agentFetch the detail (scores, grade, summary) for an agentic evaluation run.
Fetch the detail (scores, grade, summary) for an agentic evaluation run.
No parameter schema in public metadata yet.
trustmodel_guardrails_checkEvaluate a proposed agent action against the organization's cloud policy before execution.
Evaluate a proposed agent action against the organization's cloud policy before execution.
No parameter schema in public metadata yet.
trustmodel_mcp_scan_serverRun AGT's MCP security scanner over a third-party MCP server's tool list to detect tool poisoning, typosquatting, hidden instructions.
Run AGT's MCP security scanner over a third-party MCP server's tool list to detect tool poisoning, typosquatting, hidden instructions.
No parameter schema in public metadata yet.
trustmodel_shadow_discovery_scan_pathsScan local filesystem paths for unregistered/shadow AI usage.
Scan local filesystem paths for unregistered/shadow AI usage.
No parameter schema in public metadata yet.
trustmodel_shadow_discovery_fingerprint_keysDetect & fingerprint OpenAI/Anthropic API keys found on disk.
Detect & fingerprint OpenAI/Anthropic API keys found on disk.
No parameter schema in public metadata yet.
trustmodel_redteam_evaluateLaunch an adversarial red-team evaluation against a model/endpoint.
Launch an adversarial red-team evaluation against a model/endpoint.
No parameter schema in public metadata yet.
trustmodel_redteam_resultsFetch results for a red-team evaluation.
Fetch results for a red-team evaluation.
No parameter schema in public metadata yet.
trustmodel_redteam_list_probesList available red-team probes/attack categories.
List available red-team probes/attack categories.
No parameter schema in public metadata yet.
trustmodel_shadowai_scanStart a Shadow AI scan to find unregistered AI use across an environment.
Start a Shadow AI scan to find unregistered AI use across an environment.
No parameter schema in public metadata yet.
trustmodel_shadowai_resultsFetch results for a Shadow AI scan.
Fetch results for a Shadow AI scan.
No parameter schema in public metadata yet.
trustmodel_shadowai_eventsStream the detection events for a Shadow AI scan.
Stream the detection events for a Shadow AI scan.
No parameter schema in public metadata yet.
agentcert_issueMint a verifiable AgentCert for an agent off an evaluation run (signed VC + X.509). Requires an API key.
Mint a verifiable AgentCert for an agent off an evaluation run (signed VC + X.509). Requires an API key.
No parameter schema in public metadata yet.
agentcert_verifyVerify an agent's AgentCert (signature, revocation, live TrustScore, freshness). Public — no API key.
Verify an agent's AgentCert (signature, revocation, live TrustScore, freshness). Public — no API key.
No parameter schema in public metadata yet.
Sponsors
Promote your product hereReach 125k+ visitors monthlyPromote your product hereReach 125k+ visitors monthly
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any language
Agent CaffeineSponsoredClose your lid, come back to finished workSnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthlyWaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthlyknobsSponsoredSee your site on any phone, any day, in any languageConfiguration
TRUSTMODEL_API_KEYsecretTrustModel API key (free — 5 credits / $500 at https://app.trustmodel.ai) for calibrated cloud scoring. Optional: the local trustmodel_evaluate_local and trustmodel_govern tools work without it.
TRUSTMODEL_BASE_URLOverride the TrustModel API base URL (default https://api.trustmodel.ai).
TRUSTMODEL_PROFILEdefault: defaultTool profile: 'default' exposes the 7 daily-driver tools; 'security' / 'advanced' / 'all' expose all 23.
TRUSTMODEL_GUARDRAIL_FAIL_MODEdefault: fail_closedBehavior only when no HTTP response is received: fail_closed (default) or fail_open. HTTP refusals always block.
Readme
TrustModel MCP Server
A Model Context Protocol (MCP) server that lets any AI agent call TrustModel for trust evaluation, safety/bias analysis, and end-to-end agentic trace evaluation.
Works with Claude Code, Cursor, Windsurf, Claude Desktop, and any other MCP-compatible client.
One-click install
Details
- Repository
- github.com/pdxlab/trustmodel-mcp-server
- Category
- Monitoring & Observability
- Package
@trustmodel/mcp-server- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active
