Playwright

Playwright tests are the closest fit: page.goto, click, fill, and expect map one-to-one onto Navigate, Tap/Click, Type, and Assert.

Import from the dashboard

For suites that read top to bottom, skip the agent entirely: Test Cases → Import → Playwright spec, upload the .spec.ts or .js file, review the drafted steps, save. process.env.* references surface as secrets to bind to Environment values before the case is created.

Use the agent prompt below when tests lean on page objects, fixtures, test.use({ storageState }), or page.route.

Prompt

Paste this into a coding agent connected to the Autonomy MCP server from inside the repository:

Migrate the Playwright tests in this repository to Autonomy.
For each test: read it, then create one Autonomy Test Case with autonomy_test_case_create named after the test title.
Convert selectors into plain-language descriptions of the control. Keep every asserted value exactly as written.
Replace hard-coded credentials with agent-chosen fill values and configuration with environment values.
Treat page.route / waitForResponse blocks as Validate Network steps. Drop waits, retries, and framework fixtures. List existing Test Cases first and skip any that already cover the journey.
Report a table of source test -> created Test Case ID when done.

Mapping notes

• page.route() and waitForResponse() become Validate Network steps asserting on the request the UI caused.
• test.use({ storageState }) login shortcuts become explicit login steps; Autonomy authenticates for real.
• toHaveScreenshot() becomes Visual Compare against a linked Figma frame or uploaded baseline.
• Projects/devices in playwright.config become platform selection on the Test Case (web viewport presets, or iOS/Android for native).

See the migration overview for the full construct-by-construct mapping.