TestMu AI Introduces Test.md, an Agent-Native Test Framework for Kane CLI

Article content

The new framework enables replayable, human- and agent-readable tests, turning exploratory sessions into persistent, verifiable test coverage

Article content

Noida, Uttar Pradesh, May 14, 2026 (GLOBE NEWSWIRE) — TestMu AI Introduces Test.md, an Agent-Native Test Framework for Kane CLI

Article content

Article content

The new framework enables replayable, human- and agent-readable tests, turning exploratory sessions into persistent, verifiable test coverage

Article content

SAN FRANCISCO and NOIDA, India, May 14, 2026 — TestMu AI (formerly LambdaTest), the world’s first full-stack Agentic AI Quality Engineering platform, today announced the launch of Test.md, an agent-native test framework built into Kane CLI. The framework introduces a structured yet lightweight way to define, store, and replay tests, designed for both human developers and AI agents operating in modern, AI-accelerated development environments.

Article content

Article content

As software development increasingly shifts toward AI-generated code and agent-driven workflows, traditional testing frameworks are struggling to keep pace. Test.md addresses this gap by rethinking how tests are authored, executed, and maintained—eliminating the need for complex syntax, selectors, or framework-specific configurations.

Article content

Article content

At its core, Test.md enables engineering teams to convert real-time testing sessions into replayable test files using a simple markdown structure. Tests are defined as step-based objectives written in natural language, allowing them to function simultaneously as documentation and executable test cases.

Article content

“AI has fundamentally changed how software is built, but testing workflows have not evolved at the same pace,” said Asad Khan, CEO & Co-Founder at TestMu AI. “Test.md is our approach to closing that gap. It creates a shared, durable test format that both humans and AI agents can read, write, and execute, without the overhead of traditional frameworks.”

Article content

Unlike conventional approaches that rely on static scripts, Test.md introduces a replay-first execution model. Tests are authored once and then replayed deterministically, with Kane CLI intelligently deciding when to reuse recorded steps and when to re-author them based on changes in the application flow. This enables teams to scale testing without proportionally increasing effort, turning one-time validation into continuous regression coverage.

Article content

Article content

Key capabilities of Test.md include a markdown-native structure where each test is defined through step-level headings and plain English objectives, making tests inherently readable and maintainable. The framework supports modular test composition through @import, allowing teams to reuse common flows such as login, setup, and teardown across multiple tests without duplication. Configuration is embedded directly within the file using front matter, enabling environment-specific variables, runtime limits, and execution settings to be managed alongside test logic.

Article content

Test.md is designed for real-world engineering environments, with native support for CI/CD integration and agent-driven execution. Tests can run headlessly, in parallel, or in agent mode with structured outputs, making them suitable for automated pipelines. Each run generates a detailed result artifact, including step-level status, execution traces, and shareable evidence, ensuring reproducibility and traceability across environments.

Article content

A key innovation lies in its compatibility with AI-driven workflows. Test.md provides a persistent, version-controlled artifact that AI agents can generate, interpret, and reuse, enabling continuous validation of AI-generated code. This allows teams to move from “a human will verify it later” to a system where verification is built directly into the development loop.