- type: suite - title: string - description: string (supports Markdown) - tags: string[] - filter: object (or a list, treated as only) — canonical position:…
Markdown (no JavaScript)
# Reference (types) - `type:` `suite` - `title:` string - `description:` string (supports Markdown) - `tags:` string[] - `filter:` object (or a list, treated as `only`) — canonical position: after `tags:`, before `import:` / `environment:` / `items:` - `only:` string[] — run only tests/suites that have at least one of these tags (OR). Empty = all. - `skip:` string[] — do not run tests/suites that have any of these tags (OR). Empty = none. - Combined: `run iff matches only && !matches skip`. Nested running suites AND `only` lists and OR `skip` lists. - `servers:` string[] (paths to `type: server` `.mmt` files — started before tests, kept running for the suite; nested suites apply this when that suite runs) - `export:` string[] (root-only, paths to report files) - `items:` string[] (paths to `.mmt` files; use `then` to separate sequential stages; `tests` is a legacy alias) - `environment:` object (root-only) - `preset:` string - `file:` string - `variables:` object ---
For AI agents · llms.txt · Home