What problem does it solve? Long-term memory stores accumulate duplicates, stale facts, vague entries, and accidentally captured credentials over time, degrading retrieval quality. This Skill runs a structured consolidation pass that cleans, merges, and rewrites memories into durable, well-formed knowledge. ## Core Features & Use Cases - Four-Phase Consolidation Protocol: Orients by listing all memories, gathers targets via semantic search, executes deletes/merges/rewrites, then reports a summary of every action taken. - Credential and Noise Removal: Deletes API keys, tokens, raw tool output, heartbeat records, and TTL-expired operational or project memories. - Atomic Merge and Rewrite: Uses memory_update to merge duplicate facts and rewrite vague, first-person, or miscategorized entries while preserving edit history. - Use Case: After weeks of agent activity, ask for a memory cleanup to collapse 200 fragmented entries into a concise set of categorized, temporally anchored facts with zero secrets. ## Quick Start Ask the agent to run a memory consolidation pass to clean up, deduplicate, and review all stored memories.