Skip to content

ATC vs MTC

Every case Cairn designs is written as a human-readable Markdown file with a title, preconditions, steps, expected result, and recorded selectors. Cases come in two kinds:

ATC

Automatable — read-only checks on verified locators. The bot generates Playwright code for these.

MTC

Manual — form submits, security/XSS, visual judgment, irreversible actions. Left to humans on purpose.

Read-only checks (visibility, enabled state, text) on elements with reliable, verified locators — every locator is checked against the live page (getByRole().count()) before any code is written. From these the bot generates POM-style @playwright/test specs with role-based locators and test.step blocks.

Things a bot shouldn’t drive blindly:

  • full form submission with real side effects,
  • security / XSS probing,
  • visual & UX judgment calls,
  • irreversible actions.

Cairn still writes these up methodically — then deliberately leaves them to you.

If you review an MTC and decide it really is safe to automate, promote it — the case is converted to an ATC in place and picked up by the next automate run:

Terminal window
cairn promote --run runs/<id> --cases MTC-LOGIN-001

(In the TUI, press a on a case.)