Search semantic memory store for relevant past context with similarity scores.
npx skills add https://github.com/zahlmann/jarvis --skill memory-lookup
Please help me install this Agent Skill. Skill: memory-lookup Source: https://github.com/zahlmann/jarvis/tree/main/.claude/skills/memory-lookup Command: npx skills add https://github.com/zahlmann/jarvis --skill memory-lookup
This skill provides fast access to a semantic memory store at data/memories.parquet so conversations can reference past preferences, decisions, and project details, improving continuity and personalization.
Search memories related to the user's project or preferences to retrieve contextual history.
High-intent search queries and answers about installing and using this skill.
Retrieve past conversation context by querying the semantic memory store to surface relevant memories with similarity scores, timestamps, and content. Apply this in conversations where past topics, preferences, or project details need to be recalled to improve responses.
Semantic memory lookup identifies relevant memories from a semantic store to provide user context, automatically incorporating retrieved memories into responses to improve accuracy and personalization across conversations.
Search the parquet memory store by query to retrieve contextual history, returning a list of matches with similarity scores, created_at timestamps, and content to recall past project details and user preferences.
Memory lookup applies a default similarity threshold of 0.3 and ensures a minimum of three results when available, filtering the semantic store to return only relevant matches above the threshold.
Use retrieved memories to recall user preferences from earlier conversations, tailoring ongoing advice or decisions by incorporating the contextual history surfaced from the semantic store into your responses.
phuonghxStore and recall user preferences and project conventions in Markdown files.
camoaGenerate context-aware greetings based on conversational cues and time of day.
joelhooksIndex plaintext content and store embeddings for semantic search over encrypted memories.
monktuiSearch embedded documents via AgentDB vector similarity and metadata filters.
