speckit-tdd-verifylisted

# Tdd Verify Skill # TDD Verify Grade the feature's tests. Not whether they pass: whether they came first, whether they assert behavior, whether they would catch a bug, and whether every criterion in `spec.md` is actually covered. This command exists because a loop cannot grade itself. The session that wrote the tests fills in every gap from memory, so the gaps that matter are invisible to it. You read the artifacts as they stand, from cold context, and you fail closed: when the evidence is missing, the verdict is not "pass by default". You are an auditor. You do not fix what you find. ## User Input ```text $ARGUMENTS ``` You **MUST** consider the user input before proceeding (if not empty). Recognized modifiers, composable unless stated otherwise: - A feature directory name (for example `003-user-auth`): audit that feature instead of the one spec-kit currently resolves to. - `quick`: skip Phase 4 (mutation and deliberate mutants). Faster, and the verdict can be at best `PASS_WITH_GAPS`, with test strength recorded as unmeasured. - `deep`: widen Phase 4. Mutation across every file the feature touched rather than a scoped subset, and deliberate mutants on every high-risk behavior rather than a sample. - `branch`: audit everything the current branch changed, rather than one feature's test list. Use this before a pull request when the work spans features. - `--no-tasks`: write the report but do not append remediation tasks to `tasks.md`. With no input, run the