Search AI memory collections by semantic similarity with filtering and decay scoring.
npx skills add https://github.com/Hidden-History/ai-memory --skill aim-search
Please help me install this Agent Skill. Skill: aim-search Source: https://github.com/Hidden-History/ai-memory/tree/main/.claude/skills/aim-search Command: npx skills add https://github.com/Hidden-History/ai-memory --skill aim-search
This Skill allows users to efficiently search a persistent AI memory system, retrieving relevant information based on semantic similarity and advanced filtering criteria.
error_fix, rule), and intent (how, what, why).Search the memory system for information on implementing authentication.
High-intent search queries and answers about installing and using this skill.
You can search AI memory semantically to retrieve code patterns and project conventions by using vector similarity search to find information based on meaning rather than exact keyword matches.
Semantic search in an AI memory system uses vector similarity to find information based on contextual meaning, enabling you to recall past architectural decisions or error fixes without needing exact terminology.
Yes, you can filter AI memory search results by memory type like error_fix or rule, and by user intent such as how, what, or why, allowing you to narrow down specific project discussions and conventions.
You can find recently added memories in a multi-collection system by utilizing the optional decay scoring feature during vector similarity search, which boosts temporal relevance to surface newer information.
The best way to search across multiple AI memory collections at once is using cascading search capabilities, which sequentially query collections like code-patterns, conventions, and discussions to retrieve relevant results.
Yes, semantic search works for recalling past architectural reasoning by querying the discussions collection and filtering by the why intent to retrieve the historical context and project rules.
winsorllcStore and retrieve agent memory using vector embeddings and keyword search in SQLite.
rsmdtIdentify codebase patterns to enforce naming, architecture, and testing conventions.
twjarviszyra-webClassify images and match visual content to text labels.
mnsgrosaFuse vector and keyword search results using RRF, linear, cross-encoder, or cascade methods.