Workbench / Issues

Context Changes After Restore

Restoring, reopening, resuming, or reloading a task changes the context that the AI uses to continue the work.

What This Looks Like

The user restores, reopens, resumes, reloads, or returns to an AI task and the context no longer behaves the same way. Prior instructions, file references, decisions, tool state, memory, or assumptions may be missing, altered, or interpreted differently after the restore.

Why It Matters

Restored work needs continuity. If the context changes after restore, the user may believe they are continuing the same task while the AI is operating from a different state. This can cause subtle drift, broken references, repeated work, or decisions based on stale or incomplete context.

Structural Signal

A task state is expected to persist across a restore boundary, but the restored state does not match the original governing context. The issue is not just that the AI behaves differently; it is that continuity across the restore boundary is unstable.

Common Triggers

When to Use This Issue

Use this Issue when reopening, restoring, resuming, or reloading a task changes the context the AI uses to continue the work.

When Not to Use This Issue

Do not use this Issue when the user intentionally starts a new task or changes instructions after restore. Do not use it for ordinary version changes unless the restore boundary is central to the problem.

Category

Memory & Context

Primary Pattern

PAT-0190 — Contract Drift

Declared Patterns

Derived Primary Lenses

Derived Secondary Lenses

Search Intents

Ontology Metadata

Code
ISS-0031
Version
ISS-0031@0.1.0
Ontology release
0.1.0
Updated
2026-05-10T00:00:00Z

History

No public history entries recorded.

View full ontology changelog