ce-idea-scorelisted

# Idea Score **Invocation input:** Idea to score (and optional --evidence mode / --from a market-scan artifact): $ARGUMENTS Take **one** idea and produce an **opinionated, evidence-tagged verdict**: score it against a fixed seven-axis rubric with anchored 1-10 bands, run it through knockout gates and a falsifiable kill-condition, aggregate the axes by **gate-then-weight** (never a bare sum), and render a **Pursue / Pursue-with-changes / Park / Drop** recommendation. This is the **engine** the `ce-idea-scout` funnel calls to deep-dive its survivors, and it runs standalone on any single idea a human brings. > **A verdict tool, deliberately.** Unlike `/ce-market-scan` — which is evidence-bound > and structurally refuses to rank or recommend (the Scope Lock) — this skill *does* > score, weight, and recommend. That is the requested design. The discipline that > keeps it honest is therefore **not** verdict-avoidance but the **Verdict-Honesty > Contract** below: the verdict is allowed, but it may never rest on un-evidenced > guesses dressed as measurement, hide a fatal weakness behind a strong average, or > bury the per-axis shape behind one composite number. The whole artifact is labeled > **OPINIONATED decision support — a verdict, not validated fact.** > **Where it sits.** Furthest upstream, alongside `ce-idea-scout`, *before* `/ce-market-scan`: > `ce-idea-scout` (generate + triage) → human picks → **`ce-idea-score`** (deep verdict on one > idea) → human commits → `/ce-marke