Skip to Content
Agent Identity

Agents

See each person's agents and their connected instances, what each can call under your policies, and its LLM usage

When something looks wrong with an agent, you first need to know whether the problem is the agent as a whole or one connected client. SecureAuth Agent Authority keeps a separate record for each: agents (Claude Code, Cursor, Copilot, …) and agent instances (one specific connected client). The Agents tab of Inventory › Agents & NHIs lists them; for one card per agent product across everyone, see the Agent registry.

The Agents tab

The Agents tab lists one row per agent and person, for every agent a source has seen, with or without a credential. A person who reinstalls a client, or connects it from a new machine, gets a new instance of the same agent, so one row can stand for several instances. Each autonomous agent gets a row of its own, since no person owns it. Each row shows the agent, its user, its tags, the sources that saw it (the MCP gateway), its status, when a source last saw it, and when one first did. Status reads Active or Idle, and is blank for an agent no source has given a time for. Sort by Last seen or First seen. Filter by Seen in to see only the agents certain sources saw. It can also exclude sources and require all of them. For example, include Device and exclude NHI to find agents seen on a device that no NHI source lists.

Open a row to see the agent's page. Its Overview tab shows Agent Info: the agent's user, the sources that have seen it, how many credentials it has, and when it was last seen, first seen and updated. If you can view the audit log, Recent Activity for the whole agent sits beside it, with a link to its full history in Agent Trail. The Credentials tab lists every instance of the agent by its display name and ID, with its OAuth client ID, type (OAuth client or API key), status and last connection. When your console supports it, an OAuth client's row menu revokes that one instance's access. An agent with no credential has no Credentials or Effective Access tab, and no credentials count in Agent Info.

Agent page Overview tab with Agent Info beside the agent's recent activity
The Overview tab shows who the agent belongs to, where it was seen and when it was used, beside its recent activity

An autonomous agent's page shows Autonomous as its user, and its Credentials tab holds its one instance. Its Recent Activity is that instance's.

Effective access

The agent page's Effective Access tab shows which tools the agent can call on each MCP server, if you can view policies. It evaluates your policies for the instance that connected last, or the newest one if none has connected. The result is the same for every instance, unless a policy names one instance directly.

LLM Usage

When a person's agent reports OpenTelemetry, its page gets an LLM Usage tab, if you can view that person's LLM usage. It shows what the person did with the agent over the date range you pick. You see total tokens, tokens by type, tokens over time against the previous period, the top models and each session.

Agent page LLM Usage tab with token totals, tokens over time, top models and sessions
The LLM Usage tab shows one person's LLM usage of the agent

Autonomous instances

Most agent instances are user-bound: they act for the user who authorized them, so every call traces back to a person. An autonomous instance has no such user: no one authorized it, so it acts on its own behalf. SecureAuth sets the flag from whether a user authorized the instance, and you can't edit it.

  • Agents shows an Autonomous badge in place of the owner's name, in the table and on the instance's detail page.
  • To list only autonomous instances, add the Autonomous filter and choose Only. Choose Exclude to list only user-bound instances. The filter is available when you can view every user's instances.
  • In Agent Trail, the same Autonomous filter narrows events to those autonomous or user-bound instances performed.
  • In a policy condition, agent.autonomous is true for an autonomous instance. Rate limit and Data Protection conditions accept the same field.
  • In the API, each agent instance carries a read-only autonomous field, and GET /agent-instances accepts autonomous=true or autonomous=false. Audit events record the flag too, and the audit events list filters on it.

Next steps

You can now find any person's agent, see its instances, and check what it can call. Next, write a rule that applies to every instance of an agent.

  • Agent registry – one card per agent product, and slug-scoped policies
  • Policies – scope rules to an agent's slug so they apply to every instance.
  • Audit log – see the connection and activity history behind any instance.

On this page