What should an agent revisit when the world changes?
Revision-aware context snapshots, text retrieval and advisory refresh plans.
Demonstrated
Tracking revisions explicitly is enough to propose which context to refresh, with no model involved. The learned-context extension came out negative or inconclusive.
In practice
Stored context goes stale. Track what changed and plan what to re-read, instead of keeping everything.
Why this question matters
Useful context can go stale. Keeping more of it does not tell an agent which parts need another look when the world changes.
What I built
An in-memory, model-free reference for tracking revisions and planning which context to refresh.
A design choice
A change is staged as a complete snapshot and admitted only when checks pass. Until then readers see the old generation, and a failed check leaves it in place. Applying edits one at a time would be simpler, but a reader could then see half the old corpus and half the new. The cost: a candidate must cover every live document, so a removal has to be declared as a deletion.
Where the evidence stops
An in-memory reference, not a deployed model-serving or replicated distributed-memory system.