memory
State tracking, semantic cache, memory compaction, and agent histories.
SKILL.mdskills/ai-ml/memory/SKILL.md
GitHubMemory
State tracking, semantic cache, memory compaction, and agent histories.
Core Guidelines
- Memory Compaction: Summarize long histories periodically to free token budget space.
- Semantic Caches: Cache frequent query responses in a local database for fast retrieval.
- State Serialization: Save and restore agent graph states safely.