An agent needs the set of privileges sufficient for a proposed tool workflow, but a single scalar rank can obscure tradeoffs between different services, write scopes and administrator permissions. Two grants may each be minimally sufficient while exposing different resources. Selecting a broader grant because it is easier to rank loses least-privilege…
Inspect the ClaimsCase · DiggingBeagle record
Minimum agent API permissions can be incomparable rather than rankable
A study of Microsoft Graph and Google API authorization structures finds many tasks with several incomparable minimal permission sets. Frontier models often choose sufficient but dominated grants in the tested cases.
The percentages refer to tasks generated from API tables and sampled decisions; they are not rates of actual excessive permissions in production customers or known exploited OAuth grants.
Assessment profile
Assessment method- Setting
- evaluation
- Exploitation
- demonstrated
- Evidence
- Grade B
- Remediation
- not applicable
- AI role
- AGAINST AI
Basis and provenance
Enrich update: CLM-ENRICH-20261027-02 cites SRC-2026-CP21-APIAUTH-INCOMPARABLE-GRANTS at Abstract, model decisions, four risk measures and dominated grants. New evidence concerns the bounded reported mechanism, prerequisite or outcome; setting evaluation; exploit status demonstrated; remediation status not_applicable. No independently comparable impact scale, affected production population or universal exploit success measure was established, so the impact band remains null. Original Claims remain individually attributable and any observed incident or simulated result is distinguished in the narrative.
Assessed Oct 27, 2026 using diggingbeagle.assessment/1.
Evidence basis: CLM-ENRICH-20261027-02 · Incomparable Minima in Agent Tool Authorization, Xin Xu
At a glance
The case challenges one-number permission rankings, not the principle of least privilege itself.
Read the implicationsThe percentages refer to tasks generated from API tables and sampled decisions; they are not rates of actual excessive permissions in production customers or known exploited OAuth grants.
Limits and uncertaintyFull account
An agent needs the set of privileges sufficient for a proposed tool workflow, but a single scalar rank can obscure tradeoffs between different services, write scopes and administrator permissions. Two grants may each be minimally sufficient while exposing different resources. Selecting a broader grant because it is easier to rank loses least-privilege guarantees.
The paper examines 22,305 Microsoft Graph bindings and 23,852 Google API methods, reporting incomparable minimal grants on 29.8-33.3 percent of generated Graph tasks and 3.2 percent of Google tasks. The percentages refer to tasks generated from API tables and sampled decisions; they are not rates of actual excessive permissions in production customers or known exploited OAuth grants.
Use a partial-order comparison of permissions and evaluate alternative minimal grants against task-specific consequence dimensions before user consent. The case challenges one-number permission rankings, not the principle of least privilege itself. Primary evidence.
Additional evidence and trust boundary
The permission problem is a partial order, not necessarily a single ranked ladder. Where two minimum grants authorize different resources, both may be valid while exposing unequal write and administration surfaces. Consent should therefore show the set of sufficient alternatives and the concrete services, operations and administrative authority each permits. A scalar minimum can conceal this distinction. Supporting record.
Timeline
- Oct 22, 2026disclosure
Public disclosure
Claims & evidence
CLM-ENRICH-20261027-02In 1,708 modeled grant choices, the six evaluated frontier models picked a minimally sufficient option 99.0 percent of the time, but 22.1 percent of choices were dominated on every one of the authors four reported risk measures.supported
Basis: reported finding
- supportsIncomparable Minima in Agent Tool Authorization, Xin Xuprimary disclosure
Abstract, model decisions, four risk measures and dominated grants
CLM-CP21-APIAUTH-INCOMPARABLE-GRANTS-1The paper examines 22,305 Microsoft Graph bindings and 23,852 Google API methods, reporting incomparable minimal grants on 29.8-33.3 percent of generated Graph tasks and 3.2 percent of Google tasks.supported
Basis: reported finding
- supportsIncomparable Minima in Agent Tool Authorization, Xin Xuprimary disclosure
Abstract; API corpus enumeration; incomparable-minima proof and model policy experiments
Implications
The case challenges one-number permission rankings, not the principle of least privilege itself.
Controls and mitigations
- Use a partial-order comparison of permissions and evaluate alternative minimal grants against task-specific consequence dimensions before user consent.
Unknowns and contradictions
- The percentages refer to tasks generated from API tables and sampled decisions; they are not rates of actual excessive permissions in production customers or known exploited OAuth grants.
Sources and citation
Material revision history
- Oct 27, 2026 · Canonical change recorded · new in release · revision 196