Kane CLI is the terminal-native AI testing tool from TestMu AI (formerly LambdaTest). It runs from any terminal, IDE or CI pipeline — and as a first-class skill inside AI coding agents like Claude Code, Codex CLI and Gemini CLI. Describe what should happen in plain English: "log in, add a laptop to the cart, apply SAVE10, verify the total." Kane CLI drives a real Chrome browser, executes the flow step by step, and returns deterministic pass or fail with a shareable evidence link. No selectors, no boilerplate, no test framework to stand up first. Built for two audiences at once. Developers, QA engineers, PMs and vibe coders verify a feature before the PR, before the QA cycle, before deploy. AI coding agents use it as a verification primitive: they can build a feature from a prompt but cannot confirm it works in a real browser. Kane CLI closes that loop, and installs with one command. Autoheal absorbs cosmetic UI changes that break Playwright scripts. Vision-based dynamic waiting detects loaders and animations on screen — not on the network — so canvas, shadow DOM and selector-resistant frameworks work. Runs stay inside the browser viewport, taking only actions a real user could take, so nothing passes by injecting JavaScript around a validation. Resilient execution pushes through up to 50 steps per flow. Three modes cover every workflow: an interactive TUI where the browser stays open across runs, a headless CLI for scripts and pipelines, and Agent Mode with structured NDJSON output. Test.md captures any session as replayable markdown with imports and variables. Flows export to native Playwright, and existing Playwright or Selenium scripts convert in — both ways. Secrets and variables parameterize flows, custom profiles reuse authenticated sessions, and the Ask tool pauses for OTPs and CAPTCHAs. Every run syncs to Test Manager for history and proof. Install: npm install -g @testmuai/kane-cli. For agents: npx @testmuai/kane-cli-skill. Free tier included.