OpenDRS

•
3
•

Analyze pull request blast radius and assign calibrated risk scores to safely auto-approve low-risk changes into the merge queue.

OpenDRS provides an open-source implementation of Meta's Diff Risk Score (DRS) methodology, computing a calibrated 0-1 risk score for pull requests based on diff spread, git history, dependency blast radius, and author experience. Built to eliminate human review bottlenecks in fast-moving and AI-accelerated workflows, it enables automated approval and merging of verified low-risk PRs while routing complex, high-risk changes to designated human experts. Running locally without external services, it equips developers and AI agents with MCP tools, CLI commands, and GitHub Actions to assess risk, inspect hotspots, and protect production stability.

Características Principales

01Calculates a calibrated 0-1 risk score and blast radius using git history and dependency graphs
02Automates pull request approvals into CI/CD merge queues based on configurable policy gates
03Provides safety rails including file blocklists, freeze windows, size caps, and author source policies
04Offers historical calibration and backtesting commands to tune risk models to repository history
05Exposes MCP server tools for diff scoring, hotspot listing, file risk explanations, and reviewer suggestions
062 GitHub stars

Casos de Uso

01Safely auto-approving low-risk, AI-reviewed pull requests into automated merge queues
02Providing human reviewers with risk briefs, critical architectural context, and highlighted hunks
03Equipping AI coding agents with tools to analyze diff risk and check impact before opening pull requests