Stale agent authorization

An old ALLOW is reused after the task, grant, policy or resource state has changed.

01 / Attack path

How the boundary is crossed

A help-desk agent caches permission to disable account A; the incident closes before the queued action executes.

Attack path and interception point
  1. 01Lower-trust input
  2. 02Attempted instruction
  3. 03Protected action
  4. 04Enforced decision

02 / Containment

Where bounded authority limits the effect

Bind a decision to exact parameters and a short validity window; recheck active incident and grant state at the account API.

03 / Failure and evidence

The attacker’s opportunity and the defender’s record

Exploit condition

A broad cache key or absent invalidation lets an obsolete decision authorize a later effect.

Evidence to retain

Record original decision time, changed fact, cache hit, fresh denial and actual account state.

04 / Canonical scope

Why this reference stands alone

Revocation race isolates propagation timing; this page covers any stale authorization input.

Protocol or attack trace

Obsolete ALLOW test

Sequence

An account-disable decision was cached at T0; incident closes at T1; agent invokes at T2.

Negative test

T2 executor must compare task and context epochs and deny the cached verdict.

Primary references

Read the underlying material

Architecture discussion

Choose one consequential action and make its boundary explicit.

Request a Conversation