How do I add a new learning to the system?

You can use the 'add a learning' command or manually append entries to the relevant markdown file in the .automaker/memory/ directory while updating the YAML metadata.

Can I manually organize my project memory?

Yes, you can create category-specific files like patterns.md or testing.md, and use the _index.md file for human-readable documentation that isn't loaded into the agent context.

What is the Automaker memory system?

It is a system that persists project-specific learnings in markdown files within a .automaker/memory/ directory, allowing AI agents to benefit from past feature implementations.

What is the purpose of the gotchas.md file?

The gotchas.md file is used to store critical warnings and known pitfalls; it is always loaded into the agent's context regardless of score to prevent recurring bugs.

How does Claude decide which memories to load?

The system uses a weighted scoring algorithm that matches tags, relevance fields, and summaries in the memory file's YAML frontmatter against the current task description.

Automaker Memory Manager

Name: Automaker Memory Manager
Author: jbactad

byjbactad

0•

学习与文档

Manages and organizes persistent project knowledge to help AI agents learn from past feature implementations and design decisions.

The Automaker Memory Manager enables Claude to maintain a long-term memory of your project by managing structured markdown files within a dedicated directory. It utilizes a sophisticated scoring algorithm based on YAML frontmatter, tags, and importance levels to dynamically inject the most relevant context into agent prompts. By capturing architectural patterns, non-obvious constraints, and historical 'gotchas,' this skill ensures that AI agents benefit from previous work, reducing errors and maintaining consistency across the lifecycle of a codebase.

主要功能

010 GitHub stars

02Automated usage tracking including load and success rates for each memory

03Structured memory file creation with YAML metadata and markdown bodies

04Smart selection algorithm based on weighted term matching and importance

05Programmatic appending and updating of project-specific learnings

06Standardized categorization for patterns, gotchas, and API conventions

使用场景

01Building a persistent 'gotchas' list to prevent AI agents from repeating past mistakes

02Standardizing project conventions across large teams using AI-assisted development

03Capturing architectural decisions that aren't obvious from reading code alone

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jbactad/claude-plugins memory-manager

For use in Claude.ai and ChatGPT

主要功能

010 GitHub stars

02Automated usage tracking including load and success rates for each memory

03Structured memory file creation with YAML metadata and markdown bodies

04Smart selection algorithm based on weighted term matching and importance

05Programmatic appending and updating of project-specific learnings

06Standardized categorization for patterns, gotchas, and API conventions

使用场景

01Building a persistent 'gotchas' list to prevent AI agents from repeating past mistakes

02Standardizing project conventions across large teams using AI-assisted development

03Capturing architectural decisions that aren't obvious from reading code alone

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jbactad/claude-plugins memory-manager

For use in Claude.ai and ChatGPT