agent-memory

Persist repository-specific knowledge in AGENTS.md and inject it into sessions.

134|67|Updated Nov 24, 2025
One-click install
npx skills add https://github.com/OpenHands/skills --skill agent-memory-openhands
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-memory
Source: https://github.com/OpenHands/skills/tree/main/skills/agent-memory
Command: npx skills add https://github.com/OpenHands/skills --skill agent-memory-openhands

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Persist repository-specific knowledge to reduce repeated context gathering across sessions.

Core Features & Use Cases

  • Per-repo memory: Store and retrieve AGENTS.md data for each repository to guide future work.
  • Context augmentation: Automatically inject repository knowledge into your session when AGENTS.md exists.
  • Governance prompts: Prompt users to confirm what information to log, avoiding sensitive or unnecessary data.

Quick Start

Save a repository's important knowledge by creating AGENTS.md at the root (if it doesn't exist) and adding core details with user confirmation.

Frequently Asked Questions about agent-memory

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

FAQPage Schema
How do I persist repository knowledge across different sessions?

You can persist repository knowledge across sessions by maintaining an AGENTS.md file at the repository root to store details like build commands and workflows for future tasks.

How does AGENTS.md work to inject codebase context into my session?

When an AGENTS.md file exists at the repository root, its stored repository knowledge is automatically injected into your active session to provide immediate codebase context.

What is the best way to save per-repo notes like build commands and code style?

The best way to save per-repo notes is to create an AGENTS.md file at the repository root and add core details like code style and build commands to guide future work.

Can I control what information gets logged to repository memory?

Yes, governance prompts ask you to confirm exactly what information to log into the AGENTS.md file, preventing sensitive or unnecessary data from being saved to repository memory.

Does agent-memory work with any codebase for context gathering?

Yes, this approach applies across any codebase where teams save per-repo notes to reduce repeated context gathering and accelerate future tasks across sessions.

Why do I need to repeat context gathering for the same repository?

You repeat context gathering because repository-specific knowledge is not persisted by default, a problem solved by saving per-repo notes in an AGENTS.md file at the root.