councillisted

# Project Council Evaluate this project from multiple professional perspectives to surface blind spots, unknown unknowns, and cross-cutting concerns that no single role would catch alone. $ARGUMENTS ## Argument flags - `--deep` — Escalate all dispatched lenses to `model: opus` for deeper reasoning. Default is the lenses' built-in `model: sonnet`. Use when the project is high-stakes, the user explicitly asked for a thorough audit, or a previous shallow pass missed something. Cost is meaningfully higher (each lens runs on opus instead of sonnet); reserve for cases where depth is the bottleneck. Detection: if `$ARGUMENTS` contains `--deep` as a whitespace-separated standalone token (e.g. `/council security --deep`, including at the end of the args), set DEEP_MODE for the rest of this protocol. Variants like `--deep=true`, `-deep`, or `-d` are NOT recognized — use the bare `--deep` flag. - `--polish` — Narrow the lens roster to taste/excellence concerns and extend dispatch with a Jobs-grade evaluation rubric. Default lens roster becomes `visual-craft-lens` + `product-lens` + `design-lens`; security/data/sre/growth lenses are skipped unless the user named them explicitly — those audits are the wrong tool for a polish-saturated project that has already passed them. Each dispatched lens runs on opus and receives an extension prompt covering: **soul** (single-hand vs. kit-assembled feel), **signature** (one recognizable visual/interaction), **voice** (copy + tone consistency wit