Your AI coding agent says “done.” Where’s the proof? ProofShot lets agents test your UI in a real browser and show the evidence: video, screenshots, console errors and server logs. It bundles everything into a report that can go directly into a GitHub PR. Works with Claude Code, Cursor,... Codex and other shell-based agents.show more

this feels like a pretty natural next step for AI coding

This is how coding agents earn trust.

“Done” without evidence is one of the biggest weaknesses of coding agents.

Evidence is good, but I'd want it mapped to the task. Each screenshot or log should point at the acceptance criterion it proves. Otherwise you get a nice report of the happy path while the half-built feature still says done.

proofShot bundles video, logs into PR.

the "done" without receipts thing is so real like every agent i ever ran would swear up and down it worked then id open it and the button was just gone

我们团队也踩过这坑：代理报"完成"，PR 里只有 diff 没有证据，评审全靠猜。现在要求 UI 改动必须附录屏或截图，否则不合并。