A new open-source disclosure idea called “AI Code Rating” offers a short, three-character label that a project can publish in an ACR.md file. The rating is intended to help readers quickly see who maintains a repository, how much of the code was generated with AI, and how carefully AI-written changes were checked before merging.
The proposal’s core elements are consistent across outlets. The first character reflects maintainer expertise (ranging from expert to non-programmer). The second character estimates the share of code produced by AI. The third character indicates oversight level, from AI changes read and tested line by line to cases where changes are merged without review. The author emphasizes that the rating is descriptive rather than evaluative, arguing that disclosure about AI usage should be separate from judgments about software quality.
Differences in framing are mainly about audience and emphasis. Dev.to focuses on how readers otherwise must infer AI involvement from commit history and on explaining the rating’s positions in detail. Hacker News’ “Show HN” presents the specification as a simple approach to answer questions at a glance and invites community feedback, including an example repository rating and mention that the website itself follows the scheme.