agent-memory
Installation
SKILL.md
Agent memory
Memory is what lets the next conversation start from what the last ones established: the user's setup, their constraints, the decision taken last week. It fails in three ways, and every rule here answers one of them: the agent forgets what it was told, ignores a preference it holds, or holds on to what stopped being true ("still in Singapore" a month after the trip). The third is the one simple systems get worst, and the one users notice.
Apply only the routes and checks for the selected job. Read a linked skill's named file from its installed folder when available, or from the linked source. Report an unavailable required reference instead of inventing it. [skill-authoring, DISTRIBUTION.md#composing-skills]
| The job | Read |
|---|---|
| Writing memory: what to extract, who said it, grounding dates, preference atoms from parent-conversation evidence, artifact choice, duplicates, links and consolidation | WRITE.md |
| Reading it back: the always-on profile, search on demand, ranking, time-aware retrieval, token budgets, where memory goes in the prompt | READ.md |
| Keeping it true: change over time, contradictions, expiry, forgetting and erasure, what users see and correct, scopes and privacy, evals | LIFECYCLE.md |
| Compaction within one conversation, the per-turn note, history replay | the agent-harness skill |