TestMu AI launches the assurance lifecycle in Kane CLI, transforming requirements documents into verifiable test coverage
TestMu AI (formerly LambdaTest), the world’s first agent-powered AI-driven quality engineering platform, announced the “Assurance Lifecycle,” a new set of commands in Kane CLI that extends the tool’s capabilities beyond simply creating and playing browser tests. With this assurance approach, teams define the functions their product must fulfill, and Kane CLI designs tests to verify them. Each test is permanently linked to the requirement it validates, coverage is measured against verified evidence, and the test suite is updated and reconciled as the product evolves.
TestMu AI launched Kane CLI in April 2016 as a native browser-based endpoint automation tool designed for both human developers and AI-powered coding agents (including Claude Code, Codex CLI, Cursor, and Gemini CLI). By interpreting intents expressed in natural language rather than selectors, Kane CLI allows teams to describe a flow and receive a success or failure verdict accompanied by a detailed step log and screenshot, export flows to native Playwright code, and run tests headless within CI pipelines. Since its launch, it has incorporated AI-driven test case generation via the `kane-cli generate` command, which transforms a single-line description into complete test scenarios and cases. The next step is the Assurance Lifecycle: tests derived directly from actual requirements documents, not just from a prompt.
Also Read: Elastic Integrates OpenAI GPT Cyber Models into Elastic Security to Accelerate Threat Remediation
Most test suites still answer the wrong question. They report on how many tests passed, not on what requirements were verified. Tests deviate from the specifications that originated them, coverage becomes a mere estimate, and when a requirement changes, no one knows which tests have become obsolete. The assurance lifecycle addresses the pre- and post-execution stages: transforming requirements into designed tests and converting test executions into an auditable record of what has been verified.
How the insurance lifecycle works:
- Capture: kane-cli context ingest captures requirements documents (such as a PRD, specification, or policy page) and stores them in a local repository indexed by content.
- Extraction: An agent reads the sources and proposes use cases, quoting the exact lines it has read.
- Review: Teams promote proposals to a trusted status, edit them, or reject them. Everything an agent proposes starts as derived information; nothing is automatically accepted as valid.
- Design: kane-cli design tests transforms a use case into acceptance criteria, scenarios, and executable tests (exactly one test per scenario, labeled with the criterion it verifies). The resulting design is then subjected to the same review process.
- Execution: The designed tests are executed using kane-cli testmd run and kane-cli testrun run; each execution generates a certified evidence package.
- Measurement: Kane-CLI Cover reports on two axes: what an implementation has demonstrated and what remains pending according to the design. Lack of coverage is recorded as a prioritized gap, not as a lack of information.
- Maintenance: kane-cli maintain updates the test suite when a source document changes, marking as obsolete all elements extracted from the previous version until they are re-verified.
Each stage constitutes an independent command, allowing teams to pause, review, and resume the process at any time; no subsequent action is executed without prior confirmation. The assurance store resides in the project directory, operates exclusively through additions (no modifications or deletions) with a reproducible history, and remains local: source code, use cases, designs, and review verdicts are stored on the developer’s disk, not on a server. The entire lifecycle runs unattended in CI or via an AI agent; thus, a single agent can develop functionality, design tests based on specifications, and report on coverage within a single workflow.
The assurance lifecycle complements kane-cli generate rather than replacing it. Teams looking for ten good test ideas in a minute start with generate; those with a PRD who need to answer the question “what exactly is covered and how do we know?” start with assurance.
“AI agents are already writing code faster than any team can verify it, and the resulting gap constitutes a ‘validation debt.’ Kane CLI closed the first part of that gap in April by providing agents and developers with a real browser and a pass/fail verdict. The Assurance Lifecycle closes the rest,” said Mudit Singh, co-founder and chief growth officer of TestMu AI. “A test suite with positive (‘green’) results never told you what it was actually validating. Now, each test is directly linked to the requirement it verifies, each execution consolidates its evidence, and coverage becomes an auditable record, rather than a number you simply hope is correct. When a specification changes, the suite tells you which elements have become obsolete, preventing the system from silently deteriorating. That’s what assurance at the speed of autonomous agents should mean.”
The assurance lifecycle is now available in Kane CLI version 0.6.1 and later; current users can upgrade via npm or Homebrew. The extraction, design, and maintenance/reconciliation phases run on the KaneAI agent and consume credits, while the review, coverage, and storage operations are performed locally and are free.
Source: PRNewswire