challengelisted

# challenge Interrogate until it is solid — or say plainly that it already is. This skill is **advisory and write-free**: it produces findings and hand-offs, never repo artefacts, and no lifecycle transition requires a challenge pass. One stance in both modes: **do not accept vague answers; name what is missing; never invent content to fill a gap.** And the mirror rule: **"solid — nothing to add" is a first-class outcome.** An interrogator that always finds three things is noise; manufacture no objections. ## Step 0 — Preconditions and context 1. Confirm the repo is bootstrapped. Read `CONVENTIONS.md` for the record model, shapes, and language mandate; read `CONSTRAINTS.md` in full if the repo carries one — the boundary scan needs it. 2. Identify the **target**: a draft record (a proposed decision, a Draft/Agreed spec, a convention or constraint wording, a plan item) → **critique mode**; or the operator's own head ("what haven't we written down?") → **elicitation mode**. If ambiguous, ask which. ## Step 0.5 — Assessment (run first) Run the shared assessment protocol: - **Depth selector first.** Ask how deep this assessment should go: **express** — every choice takes its recommended default; only questions with no derivable default are still asked; **guided** — only the high-impact questions plus the essentials; **full** — every question. If the repo's `CONVENTIONS.md` records an `Assessment depth:`, pre-select it as the recommended option — the