trust.getModule 15Read an agent's trust score and the autonomy tier it has earned, so autonomy follows conduct the plane has actually seen.
trust.get returns an agent's trust profile: a score from 0 to 100, the number of observations behind it and when the last one landed. Released and abandoned leases, undeclared effects, ruling faults and rollbacks each move the score by a fixed amount, and time pulls it back toward 50.
Put admission in front of your MCP servers, repositories, SaaS tools and cloud accounts. Every agent gets a lease or a reason – and you keep the lever to stop any of them.
© 2026 Intramedial. All rights reserved. Intramedial is a product by Trustvice. Secured and supercharged by Avonao Labs.
Report abuseConflict frequency, policy violations and task completion are named by the product but have no safe definition yet, so they move no score. Read the score as lease hygiene, effect honesty, ruling faults and rollbacks only.
Plane module
profilenull when the plane holds none for that DID.profile.scoreprofile.samples/docs/trust-getAPI referenceRoutes, bodies and status codes./docs/trust-getOpen the consoleSee the module at work in the dashboard, on a seeded fleet or your own./dashboard-preview/trustStart freeCreate a workspace and admit your first intent./auth/signup
profile.lastObservedAtMstiertrust/tier: observe, propose, act-with-review or act.policytrust/policy: the point values, the decay and the tier floors, with the signals that are not scored.Enrolling an agent seeds its profile at 50 with no samples. Deprovisioning resets it to the same baseline.
A release adds 0.5 and an abandoned lease takes 5. An undeclared effect takes 10, an over-claimed one 2, a ruling fault 8 and a rollback an operator had to make 15.
A score decays toward 50, halving the distance every 14 days. A high score settles down without fresh good behavior, and a low one recovers without fresh faults.
trust/tier returns the highest tier whose floor the score clears: observe from 0, propose from 40, act-with-review from 65 and act from 85.
Runs in a shared demo workspace with a demo key: anyone can see what it holds, so send nothing private. trust.get docs
Press Run to send the request on the left to the shared demo workspace.
Runs in a shared demo workspace with a demo key: anyone can see what it holds, so send nothing private.
A request in the shape the API accepts, and an example response in the shape the module returns.
GET /api/v1/acp/trust/get?did=did:web:acme.example:agent:billing Authorization: Bearer av_workspace_•••• → 200 { profile }
{
"profile": {
"did": "did:web:acme.example:agent:billing",
"score": 52.5,
"samples": 7,
"lastObservedAtMs": 1793542200000
}
}Before giving an agent more to do, an operator reads its score and how many samples stand behind it. Fifty with no samples is a stranger, not a good record.
A change train reads the trust of the agents it has exposed to a change, and a missing profile holds the train rather than passing it.
A door reads a counterparty's score and sample count. No profile is reported as none, never as a neutral 50.
An agent faulted a month ago reads closer to 50 than on the day of the fault, with no operator action.
coordination.admit for that.detection.run.A module is a strong claim about a narrow thing. Rely on it for exactly what it is.
A score of 50 clears the propose floor of 40, so a new agent reads as propose, not observe.
When an observation moves an agent across a tier floor and the agent has an owner, the owner is notified, as a warning on a downgrade and as information on an upgrade.
trust/observe adds a signed number of points to a DID. It needs the policies.approve permission.
did is required for get and tier. When no profile exists, get answers profile: null.trust/policy.Authenticate with a workspace API key as a Bearer token. The base URL is https://api.intramedial.com.
/api/v1/acp/trust/getRead one agent's trust profile./api/v1/acp/trust/tierRead one agent's profile and its autonomy tier./api/v1/acp/trust/policyRead the point values, decay and tier floors.coordination.trust.getRead a trust profile as an MCP action.coordination.trust.observeRecord a manual trust observation as an MCP action.coordination.lease02Resource leases and queuesHold a resource for one agent at a time, with a lease that lapses on its own and a queue for the agents that have to wait.detection.run07Conflict detectionTurn contention between agents into a filed case with the evidence attached, instead of an alert nobody can answer.identity.enroll12Agent identity registryGive each agent its own identity before it acts, so every admitted action traces back to one agent and never to a shared account.coordination.fleetmap11Fleet MapRead every agent grant, lease, waiter and quarantine in the workspace in one call, drawn from what the plane has admitted.mcp.shadow13Shadow agent discoveryList the tool calls the MCP gateway has seen, with the agent behind each, so callers nobody enrolled show up beside the ones that are.civics.presence14Agent presenceLet agents see which peers are online and what each holds, so a quiet agent is not mistaken for a hostile one.worldmodel.graph16