Each run now seals its test definition, step-by-step screenshots, console and network logs, and failure records into a single .evidence file that teams can view, share, validate, and merge.

SAN FRANCISCO & NOIDA, India: #AgenticAI--TestMu AI (formerly LambdaTest), the Full Stack Agentic AI Quality Engineering platform, announced evidence packs for Kane CLI, its terminal-first tool for natural language browser and mobile app testing. Every Kane CLI run now produces one sealed .evidence file that holds the complete record of what the AI agent was asked to do, what it did at each step, and what happened.

Kane CLI launched in April 2026, giving developers and AI coding agents a way to run tests on real browsers and mobile apps from the terminal, in plain English, with a clear pass-or-fail verdict. Since launch, it has added AI-powered test case generation and the Assurance Lifecycle, which ties tests back to requirement documents. Evidence packs are the next step: a full record of every run.

As AI agents write and run more tests, engineering teams need a clear answer to a simple question: what exactly did the agent do, and why did the run pass or fail? That proof is often spread across screenshots, log folders, and dashboards. Evidence packs put it in one place.

Each pack contains:

• The test definition: what was asked of the agent.
• A result summary: status, duration, per-step outcomes, tags, who ran it, a share link, and the run environment. For mobile runs, this includes the device model and OS version.
• Per-step screenshots: the page as the agent saw it, plus an annotated copy that highlights the element the agent acted on.
• Browser console and network logs: captured for the whole run and tied to the step that produced them.
• Run logs: the CLI and runner logs for the execution.
• Failure records: on failed runs, a per-step record with the error message, the page state at failure, and pointers into the console and network logs.

"When an AI agent runs your tests, the first thing anyone asks is 'show me what it did.' Evidence packs answer that with one file," said Mudit Singh, Co-Founder and Head of Growth, TestMu AI. "A developer can open it, see every step the agent took, and get from a failed run to the cause in minutes. And because it is a single file, it is easy to share with a teammate or keep as a record in CI."

A pack is a self-contained zip built on the open .evidence format. Teams can work with it through three new Kane CLI commands:

• evidence serve: opens one or more sealed packs in the hosted viewer from a localhost-only server.
• evidence validate: checks a pack's integrity and completeness against a profile.
• evidence merge: combines several packs into one.

Packs also travel with the run. Runs that upload to Test Manager attach their sealed pack automatically, and replayed runs publish theirs too, so the dashboard shows execution evidence even for cache-replayed tests. For remote runs, the pack is sealed on the grid and downloaded when the job ends. If a run is interrupted, Kane CLI clears incomplete packs on its next start, with nothing to clean up by hand.

Evidence packs are available now and are produced by default on every Kane CLI run. To learn more, visit the Kane CLI evidence packs documentation.

About TestMu AI

TestMu AI (formerly LambdaTest) is a Full Stack Agentic AI Quality Engineering platform that empowers teams to test intelligently and ship faster. It offers end-to-end AI agents to plan, author, execute, and analyze software quality, enabling testing of web, mobile, and enterprise applications at any scale across real devices, real browsers, and custom real-world environments.

For more information, visit testmuai.com.

Fonte: Business Wire