Dev.to
7/2/2026

The original title is: "I built a production risk scanner in one day, here's what it caught"
Original: I built a production risk scanner in one day, here's what it caught
Short summary
Developer built BlastRadar, a tool that scores code diffs for production risk using AI analysis. It flags which systems could be affected by changes, tested on real incidents (Cursor agent data deletion, AWS outage, Claude Code wipe). Author seeks SRE/DevOps feedback on whether it's calibrated correctly and actually useful in practice.
- •BlastRadar scores code diffs on 1-10 production risk scale with plain-English explanations
- •Tested on real incidents: Cursor agent deletion, AWS Cost Explorer outage, Claude Code data wipe
- •Author explicitly uncertain about calibration and market fit; seeking direct SRE workflow feedback
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



