Kane CLI
A command-line tool that turns a plain-English objective into a real Chrome browser automation run and returns structured results, used both as a standalone browser automation CLI and as a validation layer AI coding agents call to check their own work in a real browser.
Kane CLI takes a natural-language objective, launches a locally installed Chrome browser through the DevTools protocol, drives it to completion, and returns structured NDJSON results along with an exit code, so results can be piped into other tools, run in CI, scheduled, or run on a remote grid. Coding agents such as Cursor, Claude Code, GitHub Copilot, Codex, Gemini and Antigravity can install a kane-cli skill so they call kane-cli run to verify that code they just wrote actually works in a browser before committing it.
Beyond ad hoc objectives, it also supports an assurance workflow: a context-extract command derives use cases with cited evidence from a set of requirements, a design-tests command turns those use cases into requirement-linked tests that can be committed as Markdown, and a cover command reports what is proven against what is still owed. It installs through npm, Homebrew or a shell script, and every run is uploaded to a dashboard as evidence.