TrustOps buyer walkthrough

Find server, inspect trust verdict, export policy, decide permission with /v1/decide, attach mandate, record ledger evidence, and subscribe to drift alerts.

00:00 Find server Search for io.sentinelsignal/scoring and open the canonical server detail page.
00:18 Inspect trust verdict Review production decision, freshness evidence, write-action publishing, and snapshot ID.
00:42 Export policy Generate a machine policy with allowed tools, blocked tools, scopes, and approval gates.
01:03 Decide permission with /v1/decide Ask whether this agent may call this tool, with this payload, for this user, right now.
01:24 Attach mandate Bind user authority, agent identity, budget, validity window, and tool constraints.
01:45 Record ledger evidence Persist policy snapshot, risk snapshot, mandate evidence, outcome, cost, and audit hash.
02:04 Subscribe to drift alerts Watch score, freshness, auth, schema, tool-surface, and hosted-runtime changes.

Sentinel Signal MCP

io.sentinelsignal/scoring • streamable HTTP • OAuth • 78.6 score

Open report

Inspect trust verdict

Production trust

Allow with approval

Readiness class

Safe for evaluation

Write publishing

Blocked

Fresh validation and client compatibility are visible, while write-action safeguards remain approval-gated.

Export policy

{
  "server": "io.sentinelsignal/scoring",
  "allow": true,
  "requires_human_approval": true,
  "allowed_tools": [
    "get_limits",
    "score_workflow"
  ],
  "blocked_tools": [
    "submit_feedback"
  ],
  "required_scopes": [
    "claims:read"
  ],
  "max_risk": "medium"
}

Decide permission with /v1/decide

{
  "agent_id": "agent_claims",
  "user_id": "user_ops",
  "server_id": "io.sentinelsignal/scoring",
  "tool": "score_workflow",
  "mandate_id": "mandate_abc",
  "payload_classification": [
    "phi"
  ],
  "decision": "allow",
  "ledger": {
    "invocation_id": "inv_123",
    "audit_event_hash": "5d1a..."
  }
}

Attach mandate

{
  "mandate_id": "mandate_abc",
  "scope": "score claims without submitting feedback",
  "budget_limit": 25.0,
  "valid_until": "2026-06-15T00:00:00Z",
  "constraints": {
    "allowed_tools": [
      "score_workflow"
    ],
    "blocked_tools": [
      "submit_feedback"
    ]
  }
}

Record ledger evidence

{
  "evidence_pack": "/v1/ledger/evidence-pack?server_id=io.sentinelsignal/scoring",
  "decision_replay": true,
  "policy_snapshot": "trustsnap_1e565641b01fdd41",
  "outcome": "succeeded",
  "actual_cost": 0.02
}

Subscribe to drift alerts

score drop freshness SLA new write tool auth change hosted runtime Slack webhook

00:00 / 02:40

Flow covered

Find server

Search for io.sentinelsignal/scoring and open the canonical server detail page.

Inspect trust verdict

Review production decision, freshness evidence, write-action publishing, and snapshot ID.

Export policy

Generate a machine policy with allowed tools, blocked tools, scopes, and approval gates.

Decide permission with /v1/decide

Ask whether this agent may call this tool, with this payload, for this user, right now.

Attach mandate

Bind user authority, agent identity, budget, validity window, and tool constraints.

Record ledger evidence

Persist policy snapshot, risk snapshot, mandate evidence, outcome, cost, and audit hash.

Subscribe to drift alerts

Watch score, freshness, auth, schema, tool-surface, and hosted-runtime changes.

Demo links

Start with search

Inspect trust

Compare alternatives

Machine policy

Permission gateway

Audit ledger