Workbench / Issues

Behavior Does Not Match Declared Role

The AI or agent behaves outside, below, or differently from the role, authority, responsibility, or permission posture declared for it.

What This Looks Like

The AI or agent is assigned a role, but its behavior does not match that role. It may act with more authority than declared, refuse work the role should handle, make decisions outside its responsibility, ignore role boundaries, or behave as if governed by a different role than the one presented to the user.

Why It Matters

Declared roles create expectations about authority, scope, responsibility, and limits. When behavior does not match the declared role, users cannot tell what the AI is allowed to do, what it is responsible for, or what rules actually govern its actions. This can make review and approval paths unreliable.

Structural Signal

The role declaration and observed behavior do not align. The issue is not only that the behavior is undesirable; it is that the system’s authority posture differs from the role the user was told applies.

Common Triggers

When to Use This Issue

Use this Issue when the AI or agent’s behavior does not match its declared role, responsibility, authority, or permission posture.

When Not to Use This Issue

Do not use this Issue when the role is vague and no clear mismatch can be shown. Do not use it for a single bad answer unless the behavior specifically conflicts with the declared role boundary.

Category

Permissions & Approvals

Primary Pattern

PAT-0240 — Authority-State Mismatch

Declared Patterns

Derived Primary Lenses

Derived Secondary Lenses

Search Intents

Ontology Metadata

Code
ISS-0043
Version
ISS-0043@0.1.0
Ontology release
0.1.0
Updated
2026-05-10T00:00:00Z

History

No public history entries recorded.

View full ontology changelog