memory-hygiene

1# COG Memory Hygiene Skill
2 
3## Purpose
4 
5Prevent the **stale-but-confident** failure mode: a memory or knowledge note that was correct when written ("the webhook lives at X", "the board ID is Y") silently drifts after the environment changes, yet still ranks high at recall and gets acted on.
6 
7The system move (adapted from "From Model Scaling to System Scaling: Scaling the Harness in Agentic AI", Gu, UC Berkeley, arXiv:2605.26112): **make trust a runtime decision, not a property of the stored item**. Re-verify against the live environment, and keep per-entry `last_verified` and `confidence` as first-class fields so future recalls can weigh trust.
8 
9## When to Invoke
10 
11- `/memory-hygiene`
12- "Audit my memories" / "check for stale memories"
13- After a memory misfires (a recalled fact turned out wrong)
14- Default cadence: monthly
15 
16## Scope
17 
18Sweep two stores:
19 
201. **Agent memory** — wherever your agent keeps persistent memory files (e.g. Claude Code's auto-memory directory). Sweep every entry except the index itself.
212. **Durable knowledge notes** — `05-knowledge/**` files whose claims reference the environment (paths, URLs, IDs, tool names).
22 
23A partial sweep ("just the reference entries") is fine when asked.
24 
25## Claim Classification
26 
27For each entry, split its claims into two buckets:
28 
29| Bucket | Examples | Action |
30|---|---|---|
31…