memory

tao3k's avatarfrom tao3k

The Hippocampus Interface - Vector-based Memory for LLM

⭐2stars🔀0forks📁View on GitHub🕐Updated Jan 11, 2026
[Memory and Context Management]

When & Why to Use This Skill

The Hippocampus Interface is a high-performance vector-based memory system designed for LLMs, leveraging LanceDB and FastEmbed to provide persistent, long-term context. It enables Claude to store actionable insights, recall past experiences, and maintain a dynamic knowledge base across sessions. By transforming interactions into searchable embeddings, this skill bridges the gap between short-term conversation and long-term learning, ensuring the agent evolves with the user's specific project needs and preferences.

Use Cases

  • Project-Specific Rule Persistence: Store and recall unique project constraints, such as 'all commit messages must be in English' or specific architectural patterns, ensuring the agent never forgets established guidelines.
  • Dynamic Learning & Optimization: Capture 'lessons learned' during debugging or development sessions to provide smarter, context-aware suggestions in future interactions.
  • Personalized User Preferences: Remember specific user workflows, preferred libraries, or stylistic choices to deliver a highly tailored and efficient pair-programming experience.
  • Semantic Context Retrieval: Use advanced vector search to find relevant past decisions or insights based on meaning rather than just keywords, helping to solve complex problems with historical data.
name"memory"
version"1.0.0"
description"The Hippocampus Interface - Vector-based Memory for LLM (LanceDB + FastEmbed)"
authors["omni-dev-fusion"]

Memory Skill Policy

Router Logic

Scenario 1: User wants to store something

  1. Analyze: Determine the type of memory (insight, rule, decision)
  2. Store: Call save_memory(content, metadata)
  3. Confirm: Show the saved memory ID

Scenario 2: User wants to remember/search

  1. Search: Call search_memory(query, limit)
  2. Format: Present results with relevance scores
  3. Respond: "I found X memories about that..."

Scenario 3: User asks "What have you learned?", "Show memories"

  1. List: Call get_memory_stats()
  2. Recall: Call search_memory() with relevant keywords
  3. Present: Show structured summary

Commands Reference

Command Description Example
save_memory Store insight/recipe into vector memory save_memory("Use semantic versioning", {"tag": "git"})
search_memory Semantic search in memory search_memory("git commit format", limit=5)
index_memory Optimize vector index (IVF-FLAT) index_memory()
get_memory_stats Get memory count get_memory_stats()
load_skill Load skill manifest into memory load_skill("git")

Workflow: Store an Insight

User: Remember that for this project, all commit messages must be in English.

Claude:
  1. save_memory(
       content="All commit messages must be in English only",
       metadata={"domain": "git", "source": "user"}
     )
  2. → Saved memory [a1b2c3d4]: All commit messages must be in English only
  3. → "Got it! I'll remember that commit messages must be in English."

Workflow: Recall Past Learning

User: What do we use for git tags?

Claude:
  1. search_memory("git tags semantic versioning")
  2. → Found 2 matches:
     - [Score: 0.8921] Always use semantic versioning for git tags...
     - [Score: 0.7234] v1.2.3 format for releases
  3. → "I found memories about git tags:
       - Always use semantic versioning for git tags..."

Memory vs Knowledge Skill

Aspect Memory Knowledge
Source LLM's own learnings Project documentation
Storage LanceDB (vector) File system (markdown)
Query Semantic search Keyword/pattern match
Purpose "What did I learn?" "What are the rules?"
Update Runtime accumulation Pre-indexed docs

Best Practices

  1. Store actionable insights, not obvious facts
  2. Include domain in metadata for filtering
  3. Use clear, searchable phrasing in content
  4. Recall before acting on project-specific patterns