Can I manually add entries to the memory?

Yes, you can use the command /productivity:memory add [type] to manually document a mistake or a preferred pattern.

How do I see what Claude has remembered?

You can use the /productivity:memory command to see a status overview or /productivity:memory search [query] to find specific rules.

Does it automatically learn from my feedback?

Yes, it detects natural language phrases like 'Don't do that' or 'I prefer X' to trigger a memory update workflow.

How does the memory skill store information?

It stores data locally in your project repository using markdown files such as .claude/mistakes.md, .claude/patterns.md, and CLAUDE.md.

Is my memory shared between different projects?

No, memory is project-specific as the files are stored within the project's own directory structure, ensuring context remains relevant.

Persistent Learning Memory

Name: Persistent Learning Memory
Author: goat1990

bygoat1990

生産性とワークフロー

Manages persistent project memory by documenting user corrections and preferences to prevent repetitive mistakes.

概要

The Persistent Learning Memory skill empowers Claude to grow more effective over time by maintaining a local knowledge base of project-specific rules, preferred coding patterns, and historical mistakes. By managing files such as .claude/mistakes.md and CLAUDE.md, this skill ensures that once a preference is stated or a correction is made, Claude internalizes that feedback permanently for the current workspace. It features natural language detection for automatic memory updates and a dedicated search command, making it an essential tool for maintaining high standards and consistency across complex development projects.

主な機能

Reusable pattern documentation in .claude/patterns.md with code examples
Searchable memory index to quickly retrieve stored rules and project context
Structured mistake logging in .claude/mistakes.md to prevent error repetition
Global project rule management through CLAUDE.md integration
0 GitHub stars
Automatic detection of corrections and preferences from natural conversation

ユースケース

Documenting 'lessons learned' after debugging sessions to avoid similar bugs in the future
Standardizing coding conventions for teams to ensure AI-generated code meets local style guides
Enforcing specific architectural patterns or library import styles across a codebase

概要

主な機能

Reusable pattern documentation in .claude/patterns.md with code examples
Searchable memory index to quickly retrieve stored rules and project context
Structured mistake logging in .claude/mistakes.md to prevent error repetition
Global project rule management through CLAUDE.md integration
0 GitHub stars
Automatic detection of corrections and preferences from natural conversation

ユースケース

Documenting 'lessons learned' after debugging sessions to avoid similar bugs in the future
Standardizing coding conventions for teams to ensure AI-generated code meets local style guides
Enforcing specific architectural patterns or library import styles across a codebase