Is the memory searchable?

Yes, the skill includes a dedicated search workflow using ripgrep to scan summaries, tags, and full text to find relevant memories quickly.

What is Agent Memory for Claude Code?

Agent Memory is a specialized skill that allows Claude to save, organize, and recall information across different chat sessions, creating a long-term knowledge base for your project.

Does this skill work across different repositories?

Memory is typically stored within the local project directory, ensuring that the context Claude recalls is specific and relevant to the current codebase you are working on.

How does the skill store information?

It stores information as Markdown files with YAML frontmatter in a local .claude/skills/agent-memory/memories/ directory, which can be organized into custom categories.

Agent Memory

Name: Agent Memory
Author: norifumi3

bynorifumi3

•

生産性とワークフロー

Maintains persistent knowledge and project context across multiple Claude Code sessions through a structured file-based system.

概要

Agent Memory provides a sophisticated, persistent storage solution for Claude Code, allowing the AI to retain critical insights, architectural decisions, and complex debugging findings between independent conversations. By utilizing a standardized folder structure and YAML frontmatter, it enables efficient knowledge retrieval and proactive documentation of non-obvious codebase patterns. This skill is essential for developers working on long-term projects who need their AI assistant to maintain continuity, track in-progress work, and recall historical context without re-explaining requirements.

主な機能

Categorized folder structure for organized knowledge management
Proactive logging of research findings and architectural rationale
1 GitHub stars
Metadata-rich frontmatter for advanced searching and filtering
Persistent file-based memory across multiple Claude sessions
Optimized ripgrep search workflow for rapid context retrieval

ユースケース

Building a project-specific knowledge base for team architectural decisions
Documenting complex bug fixes and non-obvious patterns to prevent regression
Tracking task state and next steps for long-running feature development

概要

主な機能

Categorized folder structure for organized knowledge management
Proactive logging of research findings and architectural rationale
1 GitHub stars
Metadata-rich frontmatter for advanced searching and filtering
Persistent file-based memory across multiple Claude sessions
Optimized ripgrep search workflow for rapid context retrieval

ユースケース

Building a project-specific knowledge base for team architectural decisions
Documenting complex bug fixes and non-obvious patterns to prevent regression
Tracking task state and next steps for long-running feature development