k2f-cv

K2F Cv — AI-Native Aesthetic Design

Purpose

Teach agents to blueprint CVs in content_and_design.md: scholarly sections, tiered long lists, citation-ready entries, and design intent—without inventing degrees, papers, grants, or metrics.

Built on K2F

This skill produces CVs as .K2F packages. Follow the k2f skill for authoring, pack/verify, and export. K2F keeps agents on sections and list roles while the engine handles multi-page flow and typography.

Before you build, install the core skill: npx skills add suzheng/k2f --skill k2f.

When to Use

• Building or updating an academic or research CV (faculty, postdoc, PhD, clinician-scientist, senior technical fellow).
• Merging exports (Scholar, BibTeX, old Word) into one structure.
• Selected vs full lists for a search—not one-page resumes (k2f-resume).

Workflow

Step 1. Generate the content_and_design.md

Single blueprint for semantic content and design intent. Use labeled blocks per section (identity, appointments, education, publications, etc.) that the k2f skill can map later.

Sub-steps

1. CV frame — Discipline, career stage, opportunity type (tenure-track, fellowship, industry lab lead), geography, and page budget (often 2–6+ pages; justify length). Note required section order if the user cites a department or funder template.

2. Input audit — Inventory appointments, degrees, publications, preprints, grants, teaching, service, talks, patents, supervision, awards, and links. Flag NEEDS USER INPUT for missing DOIs, dates, or author order; never fabricate journals, amounts, or co-authors.

3. Section architecture (include / cut / merge) — Default academic stack (reorder only when the frame demands it):

   • Identity & contact → Research interests (optional, short) → Education → Appointments / experience (reverse chronological) → Publications → Grants & funding → Teaching → Invited talks → Service & leadership → Awards → Skills / methods (sparse) → Supervision & mentoring.
   • Cut high-school, obsolete jobs unrelated to the search, duplicate talk lists, full grant narratives, hobby sections unless normative in field.
   • Merge thin consulting into appointments; split publications into Selected vs Full when lists exceed ~15–20 items or the user targets a page cap.

4. Scan layers — Glance: name, affiliation, contact, brief interests. Scan: section titles and entry leaders. Detail: dates, venues, IDs—subordinate in design intent.

5. Entry patterns — Publications: one style (authors, title, venue, year, DOI); peer-reviewed vs invited; own-name emphasis only if field-appropriate. Appointments: role, unit, dates; 0–3 impact lines with user-supplied metrics. Grants/teaching: agency, amounts, course + term when known.

6. Tailoring pass — For a stated search, elevate matching themes (methods, disease area, pedagogy) via section order, a short interests line, and Selected publications—not by adding unverified work.

7. Design intent — Tone, hierarchy, density for long documents:

   • Tone: scholarly, precise, calm; no marketing superlatives; active voice for impact lines, neutral for bibliography.
   • Hierarchy: name and current affiliation dominate; section titles clear and conventional; entry titles stronger than dates and parentheticals.
   • Density: list-forward; avoid essay blocks; publications as continuous list rhythm (hanging indent intent); whitespace between major sections, tight within homogeneous lists.
   • Palette: restrained academic; one accent; readable contrast in long lists.

Rules & constraints

• Truth gate: every degree, title, paper, grant, and number traces to user input or an explicit stub.
• Bibliography gate: do not invent DOIs, page numbers, or author lists; use metadata incomplete rather than guessing.
• Length gate: overflow → shorten via Selected lists and cut ancillary sections in the blueprint, not microscopic type.
• Consistency gate: one date format, one citation style, stable institution names across sections.
• Resume boundary: if the user needs a 1-page industry resume, redirect to k2f-resume in the blueprint note.

Common failure modes

• Resume layout on a faculty CV → scholarly sections and publication emphasis.
• Publication dump without citation rules → one style block for all entries.
• Equal weight on every item → Selected plus optional full lists.
• Missing appointment dates or degree years → halt with NEEDS USER INPUT; do not interpolate.
• Keyword stuffing in a “Skills” wall → compact methods line or categorized skills tied to real work.
• Inconsistent bolding of author name or mixed citation formats → single rule in design intent.
• Burying current affiliation below ancient history → appointments and education ordered reverse-chron within each section.

Quality checklist

• Frame: discipline, stage, opportunity, page budget, template constraints noted.
• Section stack is conventional for field; cuts and Selected vs Full choices documented.
• Glance layer shows identity, affiliation, and positioning without scrolling.
• Publications (and grants/talks) share one citation pattern; own-name emphasis rule stated.
• Appointments and education are complete on dates and institutions or flagged stubs.
• Tailoring (if any) uses reorder/selection only—no fabricated entries.
• Design intent covers tone, hierarchy, density, list rhythm, and accent for multi-page read.

Step 2. Use the content_and_design.md and k2f skill to generate a K2F file and export to an appropriate file format.

Follow k2f to instantiate sections and flowing lists from the blueprint, apply roles from design intent, verify, and export the submission-ready CV (typically PDF or DOCX).