skills.rest
dingdyan
dingdyanCOMMUNITY
Explore all skills by this author →
GitHub

agent-memory-store

Store and retrieve agent memories via a local SQLite HTTP API.

Updated Apr 5, 2026
Search Tags:#agent communication#vector search#http api#semantic memory#openai embeddings#sqlite persistence#ttl decay
One-click install
npx skills add https://github.com/dingdyan/openclaw-workspace-v2 --skill agent-memory-store
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-memory-store
Source: https://github.com/dingdyan/openclaw-workspace-v2/tree/main/skills/agent-memory-store
Command: npx skills add https://github.com/dingdyan/openclaw-workspace-v2 --skill agent-memory-store

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openai, sqlite3, python3, urllib, and includes scripts (resource) components.

What problem does it solve?

Agents often lose useful context or keep it siloed, making it hard to reuse prior decisions and knowledge across runs and across different agents.

Core Features & Use Cases

  • Cross-agent semantic memory: Store memories under an owner/agent scope and retrieve them via semantic similarity search.
  • TTL decay with persistence: Memories can expire automatically and are backed by SQLite so they survive restarts.
  • Embeddings with fallback search: Uses OpenAI embeddings when available and falls back to Jaccard similarity when OPENAI_API_KEY is not set.

Quick Start

Start the server by running "python3 scripts/memory_store.py", then send a POST to /memories with owner, content, and ttl_seconds to store your first memory.

Frequently Asked Questions about agent-memory-store

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I share semantic memory across multiple AI agents?▼

To share semantic memory across multiple AI agents, use a persistent shared memory store that supports cross-agent retrieval via semantic similarity search. Memories are stored under an owner or agent scope with SQLite persistence.

Can I retrieve agent memories using semantic vector search?▼

Yes, you can retrieve agent memories using semantic vector search. The system uses OpenAI embeddings to calculate cosine ranking for retrieval, automatically falling back to Jaccard similarity when an OpenAI API key is absent.

Do I need an OpenAI API key to use semantic memory retrieval?▼

No, you do not need an OpenAI API key to use semantic memory retrieval. The system uses OpenAI embeddings when a key is present but automatically falls back to Jaccard similarity for retrieval when the key is not set.

How do I store an AI agent memory with automatic expiration using an HTTP API?▼

To store an AI agent memory with automatic expiration, start the local HTTP API and send a POST request to the /memories endpoint. Include the owner, content, and ttl_seconds parameters to enable TTL-based decay.

What is the best way to persist shared context and knowledge across agent restarts?▼

The best way to persist shared context across agent restarts is using a semantic memory store backed by SQLite. This ensures memories survive restarts while supporting TTL-based expiration for outdated knowledge.

Related Skills

Explore More Software Engineering Skills→
aashariaashari
5

calendar-conflicts

Detect overlapping Apple Calendar events within a specified time window.

Community
Intermediate
ToowireddToowiredd
33

context-manager

Capture and recall timestamped project memories with type tagging.

Community
Intermediate
chronista-clubchronista-club
2

Qdrant スキル - クイックスタート

Enable semantic retrieval over high-dimensional embeddings using Qdrant.

Official
Advanced
javierhbrjavierhbr

sqlite-vec

Perform vector similarity search and KNN queries within SQLite databases.

Community
Intermediate