corporatelisted

# Corporate — stand up a durable-role organization Spec SSoT: `Corporate.md`. Read it before acting; this skill is the procedure, not the contract. Where the two disagree the spec wins. **Before anything else, run the rejection rubric (§15).** Two checks, in this order, and **declining is a normal, correct output of this skill** — not a failure to be worked around. *Adoption thresholds* — all three must hold: two or three duties genuinely recur *across sessions* (evidenced, not intuited); the declared resources can fund the proposed seats **at their cheapest residency**; the coordination need outlives a session. *Rejection conditions (§15.1)* — **any one firing means do not stand up an organization**: high dependency density (units of work editing interdependent parts of one codebase — the single worst-fitting case, route to one seat with context compression instead) · latency is the binding constraint on interactive work · coordination overhead exceeds the parallel gain at this scale · the available model tier is below the autonomy threshold for the duty. If any check fails, say so plainly, name which one, and recommend the cheaper shape — a Superscalar fan-out plus a board, or a single seat. Standing up an organization the resources cannot run, or that has nothing recurring to do, is the failure this step prevents. ## 1. Collect the form (§4) — four axes, with costs Ask about each; present options with their consequences rather than as a menu of words. Record the ans