claude-skill-self-improvement

🥈Silver

Analyzes Claude conversation history to identify friction patterns and generate concrete suggestions for improving your CLAUDE.md configuration and skills.

700Updated 2mo ago

Intermediate30min to implementautomation

Saves ~270 min per use

Quick InstallView Source

git clone https://github.com/bokan/claude-skill-self-improvement.git

Works with:

Claude

Overview

About This Skill

Self-Improving Claude reads your conversation history across sessions to identify friction patterns—moments where things went wrong or could work better. The skill spawns parallel agents to cross-reference these patterns against your existing CLAUDE.md and skills, then generates a ranked report with concrete suggestions for configuration updates. You review the output and apply improvements that make sense for your workflow, creating a feedback loop where Claude's performance improves over time based on real usage data.

How to Use

Clone the repository into ~/.claude/skills/self-improvement. Run /self-improvement to analyze all sessions, or use filters like /self-improvement last 3 days or /self-improvement refactoring. Review the generated CLAUDE_IMPROVEMENTS.md file and apply suggested changes to your configuration.

Use Cases

Identify recurring problems in Claude interactions and fix them systematically

Discover gaps in your CLAUDE.md configuration based on actual conversation friction

Find opportunities to create new skills by analyzing where Claude struggled

Optimize prompts and instructions by learning what already works well

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/bokan/claude-skill-self-improvement

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Analyze my previous interactions with you where I've had to correct you. Identify patterns in your mistakes and suggest improvements for [COMPANY] in the [INDUSTRY] sector. Use this data: [DATA].

Example Output

## Analysis of Interaction Patterns

After reviewing our previous conversations, I've identified several recurring patterns in my responses that led to corrections:

1. **Contextual Misinterpretation**: In 37% of corrections, I misapplied context from earlier in the conversation, particularly in technical discussions about [INDUSTRY] regulations.
2. **Ambiguity Handling**: I struggled with ambiguous queries 28% of the time, often providing overly broad answers when more specific information was requested.
3. **Data Interpretation**: When processing [DATA] sets, I occasionally misapplied statistical methods, leading to incorrect conclusions in 19% of cases.

## Suggested Improvements

1. **Contextual Priming**: Implement a system to flag and prioritize contextual cues in conversations about [INDUSTRY] topics.
2. **Ambiguity Protocols**: Develop a protocol for requesting clarification when ambiguity is detected, especially in technical queries.
3. **Data Interpretation Training**: Enhance my statistical interpretation capabilities by incorporating more advanced data analysis techniques.