skills.rest
paulingham
paulinghamCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

mcp_memory

Query long-lived recall memory via a JSON-RPC 2.0 MCP server.

1|Updated Mar 15, 2026
Search Tags:#recall#mcp#memory#server#stdio#jsonrpc
One-click install
npx skills add https://github.com/paulingham/.claude --skill mcp-memory-paulingham
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp_memory
Source: https://github.com/paulingham/.claude/tree/main/skills/mcp_memory
Command: npx skills add https://github.com/paulingham/.claude --skill mcp-memory-paulingham

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Access and query long-lived recall memory data through a stable MCP memory server.

Core Features & Use Cases

  • Long-lived stdio MCP memory server exposing recall memory via JSON-RPC 2.0.
  • Four tools: search_memory, get_timeline, get_observations, get_findings to hydrate and inspect memory data.
  • Lightweight, local-first access with privacy-conscious defaults and structured error handling.

Quick Start

Start the memory server and use the four tools to fetch observations, timelines, or scratchpad findings.

Frequently Asked Questions about mcp_memory

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

FAQPage Schema
How do I query long-lived agent memory context through an MCP server?▼

You query long-lived agent memory context through an MCP memory server using a deterministic JSON-RPC 2.0 API over stdio, enabling agents to retrieve and reason with past observations and findings.

What tools are available for retrieving recall memory data via JSON-RPC?▼

Four tools are available for retrieving recall memory data via JSON-RPC: search_memory for full-text search, get_timeline for chronological retrieval, get_observations, and get_findings to hydrate scratchpad data.

Can I use a stdio MCP memory server for local-first retrieval without external dependencies?▼

Yes, you can use this stdio MCP memory server for local-first retrieval without external dependencies, as it ships with privacy-conscious defaults and operates entirely through a local database.

How does structured error handling work in a JSON-RPC 2.0 memory server?▼

Structured error handling in a JSON-RPC 2.0 memory server follows established MCP conventions to ensure deterministic responses, preventing silent failures during memory hydration and timeline queries.

What is the best way to hydrate scratchpad findings for an AI agent?▼

The best way to hydrate scratchpad findings for an AI agent is using the get_findings tool exposed by a local MCP memory server, which fetches structured past context from a local database.

When do I need a dedicated memory server for agent context retrieval?▼

You need a dedicated memory server for agent context retrieval when your AI agents must reason with long-lived past observations, requiring structured full-text search and timeline hydration over stable JSON-RPC.

Related Skills

Explore More Software Engineering Skills→
jhl-labsjhl-labs
50

MCP (Model Context Protocol) Integration

Configure MCP servers and execute tool calls in SEPilot Desktop.

Official
Advanced
paulinghampaulingham
1

recall

Query memory.sqlite observations and scratchpad findings via Python API or CLI.

Community
Advanced
abzhawabzhaw

mcp-client-patterns

Invoke MCP server tools via HTTP JSON-RPC with discovery and error handling.

Community
Intermediate
Poorgramer-ZackPoorgramer-Zack
2

MCP Builder

Design and deploy MCP servers with typed tools in TypeScript or Python.

Community
Advanced