Agent Memory

0•

Manages a persistent, organized memory space to preserve technical insights, architectural decisions, and research findings across multiple Claude sessions.

Agent Memory provides a structured framework for Claude to store and retrieve long-term project context within a dedicated local directory. By utilizing a categorized folder structure and summary-first frontmatter, it enables efficient search and recall of complex problem-solving patterns, research discoveries, and architectural rationales that would otherwise be lost between conversations. This skill transforms Claude from a session-based assistant into a continuous collaborator that builds a project-specific knowledge base, ensuring seamless context transitions and more informed decision-making throughout the development lifecycle.

Key Features

01Summary-first search workflow utilizing high-speed ripgrep integration
02Proactive memory capture for patterns, codebase gotchas, and architectural rationale
03Persistent cross-conversation storage for research and technical findings
04Frontmatter support for tracking creation dates, status, and related source files
05Categorized folder structure with kebab-case naming for logical organization
060 GitHub stars

Use Cases

01Tracking the status and context of in-progress tasks when switching between multiple features
02Retaining complex research findings and edge-case solutions discovered during codebase exploration
03Building a local, project-specific knowledge base of architectural decisions and implementation patterns