open-mre

A LangGraph-powered tool to validate Minimal Reproducible Examples (MREs) from inbound GitHub issues.

Installation

git clone https://github.com/your-org/open-mre.git
cd open-mre

uv sync

Configuration

Set the required environment variables:

export ANTHROPIC_API_KEY="your-anthropic-api-key"
export DAYTONA_API_KEY="your-daytona-api-key"

Usage

open-mre path/to/issue.md

CLI Options

open-mre <issue_file> [options]

Arguments:
  issue_file              Path to the GitHub issue markdown file

Options:
  -o, --output-dir DIR    Directory for output files (default: current directory)
  --no-execute            Skip code execution (analysis only)
  --auto-approve-keys     Automatically approve API key usage (use with caution)

Example

# Basic usage
open-mre ./issues/bug-report-123.md

# Save outputs to a specific directory
open-mre ./issues/bug-report-123.md -o ./reports/

# Analysis only (no sandbox execution)
open-mre ./issues/bug-report-123.md --no-execute

LangGraph Local Server

Run the MRE validator as a LangGraph local server for development and testing:

uv sync --group dev
uv run langgraph dev

This starts an in-memory development server at http://localhost:2024 with:

• API endpoint: http://localhost:2024
• API docs: http://localhost:2024/docs
• LangGraph Studio UI via LangSmith

See IMPLEMENTATION_SPEC.md for detailed technical specifications.

Acknowledgements

@bracesproul for inspiring the name