BenjaminChenLab

grill-spec-tdd

Use when the user says "grill-spec-tdd", or wants to interrogate/grill requirements before a test-first delegated implementation — especially for fuzzy or high-stakes (money/auth/data) features. Triggers on requirement grilling before coding, spec-as-test, test-first-by-orchestrator, and avoiding wasted subagent runs on wrong-direction specs.

spec-tdd-coverage

Use when delegating test-first feature work to a subagent and you need EVIDENCE the new code is fully branch-covered — not just passing tests. Use when "good coverage" / "high coverage" matters, when weak or incomplete unit tests are a risk, or when spec-tdd's green-tests-but-no-coverage-evidence gap matters.

spec-tdd-escalate

Use when the user says "spec-tdd-escalate", or has a SETTLED requirement (plan/spec/design doc already locked — no grilling wanted) and wants the spec-tdd tier picked for them automatically. Triggers on requirement-already-decided, auto-pick tier, skip grilling, route to spec-tdd-lite / spec-tdd / spec-tdd-coverage / spec-tdd-adversarial.

spec-tdd-lite

Use when the user says "spec-tdd-lite", or wants test-first development WITHOUT delegating to a subagent — ONE small/medium non-critical unit (a single feature, bugfix, or refactor) in a session you'll clear after, where full spec-tdd is overkill but shallow same-context tests are still a risk. Triggers on in-session acceptance-test-first, "too small for spec-tdd", light tier, quick feature work with green-lie risk.

spec-tdd

Use when the user says "spec-tdd" or wants acceptance-test-first development delegated to a subagent — one feature, or a multi-unit batch (bug list, task-split feature). Triggers on spec-as-test, test-first-by-orchestrator, preventing weak/green-lie AI tests, circular test+implementation reasoning.

spec-tdd-task-dag

Use when driving a multi-task feature phase whose WALL-CLOCK matters and the task plan is a dependency DAG — independent tasks run as parallel waves on a task-loop substrate (authoritative plan doc + per-task commits). Parallel multiplies quota pressure ×N (cap 3, user-adjustable); the serial sibling spec-tdd-task-loop is the 429-safe mode; a plain multi-unit bug batch is NOT this skill. Triggers on parallel tasks, dependency graph, task waves, wall-clock pressure, 時段自動切換, DAG 排程.

spec-tdd-task-loop

Use when driving a whole MULTI-TASK feature phase — a task plan split into many self-contained task docs, run one task at a time with per-task commits, a plan-doc status board, and sessions that must survive the phase. The main session stays a LIGHTWEIGHT gate (compile + `git diff --stat` + JUnit-XML number recheck — never deep review, never running the tests itself); each task dispatches a level-1 sub-agent that runs the spec-tdd-escalate/tier machinery and itself dispatches the nested implementer (no self-testing — requires CLAUDE_CODE_MAX_SUBAGENT_SPAWN_DEPTH=3). Covers mock-first contract phases (the time dimension of tier choice), mid-run tier downgrades delivered by SendMessage, and resuming a half-finished task after a session break. Triggers on task loop, task-by-task spec-tdd, multi-task orchestration, per-task commit cadence, plan status board, not-yet-split requirement/blueprint needing task breakdown, 多 task 迴圈.