← ClaudeAtlas

cv-intakelisted

Parse a CV or resume that has been dropped into a person's cv/ folder and turn it into their profile.md, evidence.md and signature.md. Use when a new CV arrives, when a CV is updated, or when a person's profile still contains TBD placeholders. Run this before any email is drafted for that person.
Faaz17/Agents-and-Skills · ★ 1 · AI & Automation · score 74
Install: claude install-skill Faaz17/Agents-and-Skills
# /cv-intake, turn a CV into a usable profile Usage: `/cv-intake "<Second Person>"` or `/cv-intake "<Person Name>"` Everything downstream depends on this. A profile with `TBD` in it blocks drafting entirely, by design. ## Steps 1. **Find the CV.** Look in `<person>/cv/`. PDF, DOCX, or an image. If the folder is empty, say so and stop — do not proceed on remembered or inferred information. 2. **Read the whole thing.** Every section, including the ones that look like boilerplate. Skills lists and coursework often carry the detail that makes an email specific. 3. **Extract into structured fields:** - Full name, exactly as they write it - Degree title, university, year of study, expected graduation - GPA or CGPA and its scale, but only if they put it on the CV - Contact: email, phone, LinkedIn, GitHub, portfolio - Location and work authorisation, if stated - Every project: what it does, the stack, their role, the outcome, any link - Every internship or job: employer, title, dates, what they actually did - Awards, competitions, publications, certifications - Leadership, societies, volunteering - Skills, grouped by domain rather than in one flat list 4. **Separate hard facts from soft claims.** A CV bullet saying "improved efficiency by 40%" is a soft claim unless the person can substantiate it — mark it `(unverified from CV)`. A published paper with a venue is hard. Hard facts can go in emails; soft claims need confirmation first. 5. **Write `