From agent-memory
Use when designing memory or persistent context for an agent — deciding what to persist vs. recompute, how to retrieve only what's relevant, and how to summarize/compact history to fit the context window. Use when an agent forgets, bloats its context, or repeats work.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-memory:agent-memory-patternsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Memory is leverage, but unbounded memory is a liability — it degrades retrieval, inflates cost,
Memory is leverage, but unbounded memory is a liability — it degrades retrieval, inflates cost, and ages into wrong answers. Be deliberate about what goes in, how it comes out, and when it dies.
Out of scope: specific vector stores, embedding models, and provider context/compaction APIs — this skill is the strategy, not the storage backend.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub meaganewaller/rosetta --plugin agent-memory