Claude Reflect System enables Claude Code to learn from corrections and avoid repeating mistakes. Operations teams benefit from continuous improvement in code generation. It connects to Claude Code workflows and integrates with Python-based systems.
git clone https://github.com/haddock-development/claude-reflect-system.gitReflect System is a self-learning skill for Claude Code that creates permanent memory from user corrections. Unlike traditional AI assistants that forget between sessions, Reflect detects HIGH/MEDIUM/LOW confidence patterns from your feedback and automatically updates Claude's behavior. The system uses safe updates with timestamped backups, YAML validation, and Git integration to ensure reliability. Perfect for development teams, AI engineers, and automation workflows who want consistent code generation without repeating the same corrections. Works out of the box with Python 3.8+ and includes interactive review flows and optional auto-reflection at session end.
Copy the reflect skill directory to ~/.claude/skills/, then work normally with Claude Code. When Claude makes a mistake, correct it explicitly (e.g., 'No, use X instead of Y'), then run /reflect to analyze the session. Review the detected patterns and approve changes. Enable /reflect-on for automatic learning at session end, or use /reflect manually after each correction.
Enforce coding standards and style preferences across all Claude Code sessions
Teach Claude project-specific tools like uv, pytest, and ruff permanently
Create team learning by sharing Git-tracked skill corrections
Implement security practices and compliance checks that persist across sessions
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/haddock-development/claude-reflect-systemCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Enable the Claude Reflect System for my [COMPANY]'s [INDUSTRY] workflow. I want to integrate it with our [PYTHON_BASED_SYSTEM] to improve code generation accuracy. Please provide step-by-step instructions for setup and initial configuration.
# Claude Reflect System Integration Guide
## Step 1: Initial Setup
1. Install the Claude Reflect System package using pip:
```bash
pip install claude-reflect-system
```
2. Configure the system by editing the `config.json` file:
```json
{
"api_key": "your_api_key_here",
"learning_rate": 0.75,
"feedback_threshold": 0.5
}
```
## Step 2: Connect to Claude Code
1. Link the Reflect System to your Claude Code workflow by adding the following to your `main.py`:
```python
from claude_reflect_system import ReflectSystem
reflect = ReflectSystem(api_key='your_api_key_here')
reflect.connect_to_claude_code()
```
2. Monitor the integration status:
```python
print(reflect.connection_status)
```
## Step 3: Enable Continuous Learning
1. Activate the learning feature:
```python
reflect.enable_learning()
```
2. Set up feedback collection:
```python
reflect.setup_feedback_collection(threshold=0.5)
```AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
Automate security compliance and monitor real-time security posture seamlessly.
Automate your spreadsheet tasks with AI power
Agentic AI Workflow platform
Connected workspace for docs, wikis, and projects
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan