role-dnalisted

# Role DNA ## Overview Convert a messy job description into a **structured Role DNA**: the must-haves, weighted competencies, scoring weights, impact expectations, cultural signals, and dealbreakers that drive both sourcing (`evidence-sourcing`) and scoring (`proof-of-work-scoring`). One extraction, reused everywhere downstream. ## When to Use - A user pastes a JD or a job-posting URL and wants to source against it - Before scoring candidates — you need explicit weights and dealbreakers first - When a search is returning noise and you need sharper, prioritized criteria **Not for:** running the search itself (`evidence-sourcing`) or evaluating a candidate (`proof-of-work-scoring`). ## Input - **JD text** → extract directly. - **JD URL** → fetch first. Use the host `scrape_job_description` recruiting tool if present, else host **Exa** `web_fetch`/web_search. Never invent requirements that aren't in the source. ## Role DNA Schema Produce this object: ```json { "title": "Staff Backend Engineer", "seniority": "staff", "location": "Remote (US) | NYC | …", "must_have_requirements": [ { "requirement": "7+ yrs building distributed systems", "criticality": "CRITICAL" }, { "requirement": "Production Go or Rust", "criticality": "CRITICAL" }, { "requirement": "Owned a system at scale (>10k rps)", "criticality": "IMPORTANT" } ], "core_competencies": [ { "name": "Distributed systems depth", "weight": "CRITICAL" }, { "name": "Code quality / OSS sig