Tools: Github Browser Plugin For AI Contribution Blame In Pull Requests

With the proliferation of effortless code‑generating tools like Claude Code, Codex, and Cursor, slop‑slung contributions are being doled out as outright spam in hopes of getting a name tacked onto popular open‑source projects. Most are well‑intentioned — it’s just that this workflow is entirely new, and the tools and norms haven’t been established yet. Some open‑source projects have publicly banned them (see: zig, tldr, ghostty), going so far as to vet contributors into a select trusted group.

Oftentimes, depending on the preference of the team and project, less consequential and isolated code could warrant a 100% AI contribution. Non‑user‑facing tooling, a private beta feature, or a proof‑of‑concept immediately come to mind. The ability to retroactively see which parts of the codebase were AI contributions, especially in these use cases, could be very valuable. What was tabbed in by Cursor at 3am six months ago could be a part of today’s refactor.

Projects like Zig may never allow ai contributions, and I am not here to argue that they should change this stance. But in other cases, where the reaction is a heavy‑handed outright refusal, maintainers and developers could have a change of heart if they could codify an allowable percentage done by AI in each pull request. Even without a hard‑and‑fast rule, a percentage could serve as a sort of gut check — an overall score as part of a bigger picture of quality in a PR.

The git-ai project allows you to automatically track agentic ai code contributions across your team and codebase, zeroing in line-by-line, preserving code-generating prompts, all while working within common git workflows. Git-ai works by extending and enhancing your current tooling without slowdown (thanks to Rust) while ‘staying out of the way’ - so you can work as if it’s not even there.

git-ai stores things like per-line ai contributions, the model and prompt given for the code generated.

git-ai works by storing this ai contribution data in git notes. Git notes are simply blobs attached to commit refs. It’s eloquent in that the meta data stays with the commit, git-ai also contains additional instrumentation to “survive a merge --squash, rebase, reset, cherry-pick etc.”

🤖 Track AI code in a Multi-Agent world. Because developers get to choose their tools, engineering teams need a vendor agnostic way to track AI impact in their repos.

🎯 Accurate attribution from Laptop → Pull Request → Merged. Claude Code, Cursor and Copilot cannot track

Source: HackerNews