Does Agent Memory store my bugs and error logs?

No. The skill is specifically instructed to ignore issue-specific information and only log general knowledge that remains helpful across different future tasks.

Is the AGENTS.md file loaded automatically?

Yes, if the file exists in the repository root, Claude will automatically add its contents to your context when you start a session.

What is the purpose of the AGENTS.md file?

AGENTS.md is a markdown file stored in your repository root that Claude uses to store and retrieve long-term knowledge, such as build commands and directory structures, specific to your project.

How do I manually save information to the memory?

You can use the /remember trigger to prompt Claude to summarize its recent observations and suggest items to save to the AGENTS.md file.

Can I review what Claude writes to the memory file?

Yes, Claude is required to ask for your confirmation by listing the exact items it plans to save, and it will only proceed with the items you approve.

Agent Memory

Name: Agent Memory
Author: plurigrid

byplurigrid

•

Productivity & Workflow

Persists repository-specific context, workflows, and project standards across Claude Code sessions using a centralized AGENTS.md file.

Agent Memory enhances Claude's long-term utility within a codebase by maintaining a living document of project-specific knowledge. It automatically captures and stores essential information such as build commands, directory structures, preferred coding styles, and common developer workflows. By leveraging an AGENTS.md file in the repository root, this skill ensures that Claude 'remembers' the unique characteristics and requirements of your project between sessions, drastically reducing the need for repetitive onboarding and ensuring consistent adherence to project standards.

Key Features

01Structured organization of build, lint, and test commands for future use

02Automatic loading of repository-specific metadata into active sessions

038 GitHub stars

04Interactive memory capture with mandatory user confirmation for all updates

05Persistent context storage via AGENTS.md files in the repository root

06Dynamic repository structure documentation and architectural mapping

Use Cases

01Automating the onboarding process for Claude Code in complex or legacy repositories

02Standardizing project-specific commands to prevent execution errors in future sessions

03Maintaining a consistent style guide and workflow reference for AI-assisted development

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi agent-memory

For use in Claude.ai and ChatGPT

Download Skill