Privilege AIMKC2 results ↗

AI agent permissions

Every agent permission question starts with identity. An agent acting as itself is a new privilege; an agent acting as its user is a new interface.

READ THE DEFINITION ↓IDENTITY · SCOPE · EXPIRY

01 / DEFINITION

The short answer first.

What are AI agent permissions?

AI agent permissions are the rights an agent holds when it calls tools and reads data: which identity it acts as, which resources that identity may reach, which operations it may perform, and for how long. Permission design sets the ceiling on everything an agent can do, correctly or otherwise.

Permissions and access control are often discussed as one thing and are better separated. Permissions are the design: what rights should exist and who holds them. Access control is the enforcement: how each request is checked. A clean permission model badly enforced is a document, and strong enforcement of an over-broad model is a well-guarded mistake.

WORKED EXAMPLE

Map permission to a calendar operation

A meeting assistant needs to find open times and draft invitations. Reading availability, reading private event titles and sending an invitation to an external guest are different operations. Granting calendar access as a single permission can expose more information or action authority than the task requires. Start with the smallest operation the product actually needs.

Trace delegated authority through the request

When a user asks for an available slot, pass the user's identity and the permitted calendar identifiers into the scheduling service. Let that service decide what information may be returned. A broad service credential can authenticate the application while still requiring an individual authorisation check. Possession of the application's credential does not establish the user's right to read a particular calendar.

Create tests for the user's own calendar, a colleague who shares availability only and a calendar that is not shared at all. The assistant should return a useful open slot without exposing a private meeting title. Then ask it to invite an external address. If sending requires confirmation, the application should bind approval to the recipient, time and meeting details rather than granting open-ended permission to send.

Finally revoke the calendar share while a proposed invitation is waiting. Check both a fresh request and a queued operation. An expired permission must stop the action at execution, even if a previous model turn correctly observed the slot. Save the effective identity and decision reason so support can distinguish an expired share from a model-planning error.

02 / THREE IDENTITY MODELS

Identity models for agents. The choice determines whether the agent can become a privilege amplifier.
ModelHow it worksUse whenRisk
Agent as itselfA dedicated service account with its own rightsBackground work with no requesting userAggregates privilege; attribution is lost
Agent as the userThe requesting user's identity is propagated to every callAnything triggered by a personRequires identity propagation through every tool
Agent with delegated scopeA short-lived token carrying a subset of the user's rightsLong-running or asynchronous tasksToken lifetime and scope must both be bounded

The default should be the second, with the third for work that outlives the session. The first is common because it is easy, and it is how an agent becomes the most privileged actor in an organisation: a service account accumulates the union of every capability any user ever needed, and then a system that follows instructions from documents holds all of it.

Identity propagation has a real engineering cost. Every tool has to accept and honour a caller identity, which for internal services is straightforward and for third-party APIs is frequently impossible. Where it cannot be propagated, the compensating control is a narrow, purpose-specific credential per tool rather than one broad account shared across them.

03 / SCOPING TO THE TASK

01

Start from the task, not the role.

Enumerate the calls this task requires. Role-based grants imported from human job descriptions are far broader than any single task needs.

02

Bind the grant to the request.

A token issued for this run, this user and this objective, expiring when the run ends rather than on a calendar.

03

Separate read from write from send.

Three different risk profiles. Combining them into one grant removes the ability to allow the safe part of a task.

04

Cap the volume.

A limit on records touched per run. Legitimate access to one matter is not legitimate access to the whole corpus.

Volume caps are underrated. Most permission models are binary — an identity may read a collection or it may not — which makes a compromised agent's access identical to a legitimate one's right up to the point where it has read everything. A per-run cap converts a silent bulk read into a visible boundary event, and the cap rarely inconveniences real work because real tasks touch few records.

04 / WHAT GOES WRONG

↳ 01

Permission creep.

Capabilities added during development to unblock a demo are never removed. Expiring grants by default is the only reliable countermeasure.

↳ 02

The union problem.

Several narrow tools combine into a capability nobody approved. Review the set, not the items — see privilege escalation.

↳ 03

Retrieval as a side door.

An index built without document-level permissions hands the agent everything, whatever its API rights say. See RAG access control.

The third pattern is the one that most often defeats an otherwise careful design. Permissions are enforced at the API and then quietly bypassed by a vector index built from the whole corpus with no per-document access metadata. Retrieval is an access path and has to be permissioned as one. Privilege AI treats permissions and retrieval boundaries as the same problem: what information a model can access, and which actions it can propose.

05 / QUESTIONS

Should an AI agent have its own account?

Only for work with no requesting user, and then with a narrow, purpose-specific grant. For anything a person initiates, propagating that person's identity keeps existing access controls meaningful and preserves attribution in the audit trail.

How do you scope permissions for an agent that explores?

Scope the resources rather than the sequence. An agent that may need any of a hundred records within one matter should be permitted that matter and capped on volume, not granted the whole store because its path is unpredictable.

What about agents that call other agents?

Delegation must narrow, never widen. A sub-agent inherits at most the caller's scope, with its own shorter lifetime, and the chain is recorded so the original requester remains identifiable. Chains that re-broaden scope are a common escalation path.

How long should an agent credential live?

As long as the task, which is usually minutes. Long-lived credentials exist for operational convenience and are the main reason a compromise persists beyond the run that caused it.

Do permissions need to cover retrieval as well as tools?

Yes, and it is the most frequently missed half. A retrieval index built without per-document permissions gives an agent access it was never granted through any API. See RAG access control.

WORK WITH PRIVILEGE AI

How we can help

Talk with Privilege AI

Tell us which identities and tools your agent uses, and which permission you would like to narrow or verify.

Talk to us →

CONTACT PRIVILEGE AI

Contact Us

Tell us which identities and tools your agent uses, and which permission you would like to narrow or verify.

When should we call?

By submitting, you agree that Privilege AI, a division of Tech1e LLC, may call or email you about this inquiry. Please leave confidential information out of your message.

Give it
only the task.

Privilege AI designs permission and retrieval boundaries that define what a model can access and which actions it can propose.