Docket gives every AI-written commit a receipt for what was never reviewed
Docket – Per-commit evidence records for agent-written code
Coding agents produce diffs faster than anyone can check them, and the implementation journey—what the agent tried, what it verified, what nobody looked at—is thrown away at commit time. Docket captures that journey from Claude Code, Codex CLI, and opencode transcripts, folds it against the diff, and writes a signed per-hunk evidence record on an orphan ref. No account, no network. `docket show HEAD` surfaces what was tried, what failed, and which lines have no human contact.
An attribution engine that is confidently wrong is worse than no product, so `unknown` is always an available answer, and it always comes with a reason.
- sleeby
All the AI tools usually add "Co-Authored-By XXX" as a trailer on commits where an agent did most of the work.
Cheap and it survives every git tool.
What does this give me beyond that?
- Yukonv
Interesting approach to documenting an audit record of changes instead of dumping info into commit description that clutters history. The one major risk not addressed by this vibe code is orphaned references are not a safe storage location. A `git prune` will delete them all since they are not part any head reference. https://git-scm.com/docs/git-prune
- bastawhiz
Who is this for? Who's looking at it? It seems like this is solving an engineering culture/process failure by introducing new engineering process. Like I'm trying to understand the circumstances where a) I don't have CI/mandatory reviews/PRs to organize changesets with things like intent and b) even though these commits are allowed to make it to main, I want to know this information. Is it to blame coworkers when things go sideways? To avoid deploying infra?