tailor-cvlisted

# tailor-cv Generates a per-job tailored CV for every role marked `selected: true` in the **Role Pipeline** artifact's `jobs` database collection that doesn't already have one. This is Phase 2: **CV tailoring only** — no application filling happens here (that's Phase 3). The sample template (`cv-template.tex`) has two experience entries — a current role and an earlier role — matching the two fact banks below. **Nothing in Experience or Projects is fixed copy** — `work-experience-bank.json`, `undergrad-experience-bank.json`, and `project-bank.json` all hold *facts* (what actually happened, in plain terms, tagged with tech), not final copy. Tailoring a CV means: read the job's title/fitNote/description, decide which facts are relevant, and **write fresh bullets in the user's voice that foreground the angle this specific job cares about** — e.g. a data engineering role gets pipeline/scale-relevant facts written up in that language; a full-stack role gets the same underlying project's frontend facts written up in delivery-and-frontend language. Two bullets drawing on the same fact for two different jobs should read differently. **The two experience roles are selected jointly, not independently.** Not every job needs every fact from both banks — pick whichever facts from *either* bank best fit this job, across both roles at once, the same way you'd pick projects. When a current-role fact and an earlier-role fact cover genuinely overlapping ground, **the current-role one wins**