skills.rest
midnghtsapphire
midnghtsapphireCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

GBrain — Persistent Agent Memory

Store and retrieve persistent agent memory using Markdown files and vector embeddings.

1|3|Updated Feb 20, 2026
Search Tags:#memory#persistence#pgvector#knowledge-base#agent-memory#brain
One-click install
npx skills add https://github.com/midnghtsapphire/revvel-standards --skill gbrain-persistent-agent-memory
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: GBrain — Persistent Agent Memory
Source: https://github.com/midnghtsapphire/revvel-standards/tree/main/skills/gbrain
Command: npx skills add https://github.com/midnghtsapphire/revvel-standards --skill gbrain-persistent-agent-memory

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

AI agents start new conversations with amnesia, losing context about people, deals, and ideas. GBrain provides a persistent, searchable memory that is loaded into context before every response and updated after each interaction.

Core Features & Use Cases

  • Persistent knowledge base built from plain Markdown files that humans can edit.
  • Retrieval-augmented memory with fast search and vector embeddings to surface relevant memories.
  • Use Case: An agent recalls last meeting notes, open tasks, and deals from prior sessions to inform current decisions.

Quick Start

Install Bun, install GBrain globally, initialize the brain, and connect it to your AI agent to begin recording and retrieving memory.

Frequently Asked Questions about GBrain — Persistent Agent Memory

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

FAQPage Schema
How do I give my AI agent persistent memory across different sessions?▼

You can establish persistent agent memory by storing learned context in a lightweight knowledge base of plain Markdown files. The agent reads relevant memories before responding and writes new context after interactions.

How does vector search work for retrieving agent memory?▼

Vector search retrieves agent memory using local indexing with PGLite and pgvector. This retrieval-augged approach surfaces relevant memories by calculating vector embeddings for fast contextual search.

Do I need Bun to set up a local knowledge base for AI context?▼

Yes, you need to install Bun to set up this local knowledge base. You must install the tool globally, initialize the brain repository, and connect it to your AI agent to begin recording memory.

Can I manually edit the knowledge base files used for agent memory?▼

Yes, the persistent knowledge base is built from plain Markdown files that humans can edit. This allows you to directly manage and update the stored context your AI agent retrieves.

What is the best way to maintain long-running context across conversations and projects?▼

The best way to maintain long-running context is a memory-loop workflow that ensures deterministic memory growth. This workflow reads context before responses and writes updates after each interaction.

Related Skills

Explore More Software Engineering Skills→
cloudofgeorgecloudofgeorge

setup-gbrain

Install the gbrain CLI, initialize a PGLite or Supabase brain, and register an MCP.

Community
Advanced
mengsi16mengsi16
1

brain-base

Query and persist local documents through a structured knowledge base workflow.

Community
Intermediate
mindmorassmindmorass
2

qdrant-patterns

Store and retrieve documents in Qdrant for semantic search.

Community
Advanced
InverterNetworkInverterNetwork

pinecone-research

Store and retrieve conversation context using Pinecone vector databases.

Official
Intermediate