What problem does it solve? Agents lose context between sessions and cannot reliably recall past decisions, notes, or project knowledge. This Skill provides a persistent, searchable memory layer so the agent can retrieve prior context, maintain daily logs, and keep a curated knowledgebase without manual file digging. ## Core Features & Use Cases - Hybrid Memory Search: Query daily logs, knowledgebase, and weekly reports using BM25 keyword search combined with local vector embeddings via the qmd-query MCP tool or hybrid-search.py fallback. - Memory Maintenance: Update MEMORY.md with permanent facts, append daily context logs, and organize long-term knowledge into structured knowledgebase folders. - Session-End Memory Flush: Automatically save decisions, action items, and summaries before a session ends, then re-index content for future retrieval. - Use Case: Ask "what did we decide about the firmware timeline last month?" and the agent searches indexed daily logs and PEC minutes, returning cited source files with the answer. ## Quick Start Ask the agent to search memory for a past decision or topic, such as "find what we know about the project alpha release timeline."
