Stats
Actions
Tags
From claude-memory
This skill should be used when the user asks to search, query, or recall project memories, observations, or past session history. Trigger phrases include "what do you remember about", "search memory for", "recall", "what did we decide about", "find past observations on", "check memory for", "show me what we learned about". Also triggered by /memory-query.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-memory:memory-queryThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Search the project memory database for relevant observations and insights.
Search the project memory database for relevant observations and insights.
Execute the query script using the production venv with system fallback:
~/.config/claude-memory/.venv/bin/python3 ${CLAUDE_PLUGIN_ROOT}/scripts/query.py "{{query}}" 2>/dev/null || python3 ${CLAUDE_PLUGIN_ROOT}/scripts/query.py "{{query}}"
Results are ranked by a composite relevance score (0.00–1.00) with these weights:
Each result line follows one of these formats:
[0.XX] INSIGHT: ... — A consolidation insight (synthesized from multiple observations, boosted priority)[0.XX] P{N}: ... — An individual observation with priority level P1–P4
Topics: ... — Topic tags for the preceding observation Summary: ... — Summary text for consolidation insights~/.claude/projects/*/claude-memory.db or at $CLAUDE_MEMORY_DB_PATH.If no results are found for the query:
After presenting results, offer to:
/memory-status to check database health and observation counts/memory-consolidate to synthesize unconsolidated observations into insightsnpx claudepluginhub seangsisg/deja-claudeSearches persistent cross-session memory database (claude-mem) to retrieve past work, decisions, and context. Use when user asks about prior sessions or solutions.
Implements 3-layer memory search workflow to recall past work, decisions, errors, and project history token-efficiently via layered functions.
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.