speckit-tdd-runlisted

# Tdd Run Skill # TDD Run Drive the loop. One behavior from the test list, one failing test, the smallest change that makes it pass, a refactor step while green, and an evidence entry in the cycle log. Then the next behavior. This is the only command in this extension that writes tests and source code. Besides those it writes only the feature's `tdd/test-list.md` and `tdd/cycle-log.md`, and the checkboxes of the `tasks.md` tasks it completed. It does so under a discipline that is the whole point: **the test is written first, it is observed failing, and the failure output is recorded before the implementation exists.** A cycle without recorded red evidence is a cycle that proved nothing, and `/speckit.tdd.verify` will say so from cold context later. The pressure in this command is always toward doing more per step: write the test and the code together, take a big step because the answer is obvious, fix a second thing while you are in the file. Resist all of it. Small steps are not a formality; they are what makes each green mean something specific. ## User Input ```text $ARGUMENTS ``` You **MUST** consider the user input before proceeding (if not empty). Recognized modifiers, composable unless stated otherwise: - A behavior id (for example `U3`, or `U3 U4 U5`): run the loop on exactly those behaviors, in the order given, instead of the next pending one. - `all` (the default): keep cycling until every behavior is `DONE`, the list is blocked, or an escape hatch fires