- Home
- Skills
- Agents & Orchestration
- Brain — long-term memory for coding agents
More screenshots
Works with the AI Agents you already use
Brain — long-term memory for coding agents
Local long-term memory and multi-agent coordination protocol for persistent coding context.
$10
About this skill
Coding agents lose context when a session ends or when switching between different tools. Important architectural decisions, hard-won bug fixes, and project-specific preferences vanish, forcing you to repeat instructions. This package provides a durable, local long-term memory layer using the Brain MCP server to ensure knowledge persists across sessions and agent handoffs.
What it does
- Persistent memory storage saves decisions, snippets, and todos in a local database that survives session restarts.
- Multi-agent coordination enables messaging and handoffs between different agents like Claude and Codex using a formal protocol.
- Context retrieval searches project-specific or global knowledge bases to prevent contradicting prior decisions.
- Knowledge hygiene audits existing memories for duplicates, stale dates, or oversized entries to maintain context quality.
- Automated upkeep proposes merges and updates based on local evidence while protecting sensitive record types.
How it works
- Setup installs the
agent-memory-brainserver viauvorpipxand registers it with your local agent host. - Capture uses
memory_saveto store actionable facts as specific kinds likedecision,preference, orsnippet. - Query calls
memory_searchorproject_contextat the start of tasks to pull relevant historical data. - Coordinate sends messages between agents via
message_postand tracks replies withmessage_waitfor complex workflows.
Frameworks & tools
This skill requires Python 3.12+ and the Brain MCP server. It is designed for use with Claude Code, Cursor, or any MCP-compliant agent host. It uses uv for tool management and stores data in a local SQLite database.
Why this beats prompting it yourself
Standard prompting relies on a limited context window that clears every session. This skill implements a structured retrieval-augmented memory system that survives reboots. It includes built-in deduplication and relationship mapping that raw system prompts cannot maintain reliably.
Use cases
- Record why a specific library was chosen to prevent an agent from suggesting a migration later.
- Pass a task state and context from a research agent to a coding agent via the handoff protocol.
- Store a global preference for small PRs that applies across all your local repositories.
- Audit a project knowledge base to find and merge overlapping architectural notes.
Known limitations
Memory content is capped at 2,000 characters by default. It does not store secrets or credentials. Requires a local Python environment for the MCP server.
Content capped at 2,000 chars per entry. No storage of secrets/tokens. Requires Python 3.12+. Multi-agent handoffs require all agents to support the Brain protocol.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
Requires
Permissions
Allowed Hosts
File Scopes