Time-of-check to time-of-use authorization failure

The protected request changes between permission evaluation and execution.

01 / Attack path

How the boundary is crossed

An agent obtains approval for deployment to staging, then a mutable environment alias points to production when the executor runs.

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

Resolve immutable target identity and normalized parameters at execution, bind them to the decision and reject drift.

03 / Failure and evidence

The attacker’s opportunity and the defender’s record

Exploit condition

The decision is correct for the checked object but execution affects a different object.

Evidence to retain

Retain checked resource ID, executed resource ID, binding result and observed deployment target.

04 / Canonical scope

Why this reference stands alone

Stale authorization concerns old state; this page owns a changed object or argument after a valid check.

Protocol or attack trace

Target-switch race

Sequence

Policy approves deploy to alias staging, then alias points to production before commit.

Negative test

Bind immutable target ID and payload digest at executor; deny alias or payload drift.

Primary references

Read the underlying material

Architecture discussion

Choose one consequential action and make its boundary explicit.

Request a Conversation