Can I manually add items to Claude's memory?

Yes, you can use the command '/productivity:memory add [type]' to manually document a mistake or a pattern without waiting for Claude to detect it automatically.

How does the memory skill learn my preferences?

The skill monitors your conversation for natural language corrections, such as 'No, do it this way' or 'Always use X instead of Y', and automatically updates the relevant memory files.

How do I see what Claude has learned so far?

You can run the '/productivity:memory' command to see a summary of all stored rules, errors, and patterns, or use '/productivity:memory search' to find something specific.

Which files does this skill use to store information?

It manages three primary files: .claude/mistakes.md for avoiding past errors, .claude/patterns.md for preferred code structures, and CLAUDE.md for high-level project rules.

Persistent Memory Manager

Name: Persistent Memory Manager
Author: goat1990

bygoat1990

Productivity & Workflow

Enables Claude to learn from mistakes and store persistent user preferences to ensure consistent coding behavior across sessions.

About

This skill provides a robust framework for Claude to maintain a long-term memory of project-specific preferences, recurring mistakes, and preferred implementation patterns. By automatically detecting user corrections and maintaining structured files like mistakes.md and patterns.md, it transforms Claude from a stateless assistant into a context-aware collaborator that continuously improves based on your feedback. It serves as a personal knowledge base that prevents the repetition of previous errors and strictly adheres to your specific architectural style and coding standards.

Key Features

Pattern library management for reusable, project-specific code snippets
0 GitHub stars
Automatic correction detection from natural language feedback
Global rule enforcement through CLAUDE.md integration
Structured mistake tracking via mistakes.md to prevent error regression
Interactive memory search and manual entry commands

Use Cases

Enforcing project-wide coding standards and specific library import preferences
Maintaining a 'living' documentation of architectural decisions and preferred design patterns
Reducing debugging time by ensuring Claude remembers previous project-specific fixes

About

Key Features

Pattern library management for reusable, project-specific code snippets
0 GitHub stars
Automatic correction detection from natural language feedback
Global rule enforcement through CLAUDE.md integration
Structured mistake tracking via mistakes.md to prevent error regression
Interactive memory search and manual entry commands

Use Cases

Enforcing project-wide coding standards and specific library import preferences
Maintaining a 'living' documentation of architectural decisions and preferred design patterns
Reducing debugging time by ensuring Claude remembers previous project-specific fixes