Show HN: Model Due Diligence

Category: security

Tags: ai-safety, supply-chain-security, static-analysis

Score: 7.0/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 6)

Model Due Diligence is a static supply-chain security tool for AI model files and repositories, detecting unsafe serialization, secrets, malicious code patterns, and provenance issues before runtime import. Its innovative combination of native scanners, external tool adapters, and risk scoring fills a practical gap for AI safety, though it is still early-stage with no stars.

Target audience: security engineers, ml engineers, devops

Repository: https://github.com/mmccalla/model-due-diligence · Python · Apache-2.0

View on Hacker News