cv-building

End-to-end CV building pipeline as an AI agent skill. Works with Claude Code, Cursor, Windsurf, Codex, or any agent that supports skills. Semantically accumulate career data into a master doc (premix storage), pour targeted molds (cv.yaml), render professional PDFs via RenderCV, and produce strategic interview materials.

Philosophy: Premix Storage & Molds

• Master doc (master.md) = premix storage — neutral, ever-growing, infinite potential
• CV molds (cv_<company>_<position>.yaml) = specific shapes poured from the premix
• Same master, many molds — one biography, many tailored CVs for different targets

Pipeline

Any Input → Semantic Accumulation → Master (premix) → Mold (cv.yaml) → Package
     ↓              ↓                    ↓                  ↓              ↓
Free-form     Merge into          Neutral, growing    Tailored for    PDFs + strategic
CV/fractals/  storage              career storage       target role    materials
conversation

4-Step Process

1. Semantic Accumulation — Accept any input format, merge into cv_files/master.md, track in cv_files/CHANGELOG.md
2. Master Normalizing — Organize raw data into neutral template (no tailoring at this layer)
3. Quality Gate Processing — Pour a mold (cv_<company>_<position>.yaml) with industry-appropriate tone:
   • STEM (engineering, research): Strict anti-cringe, metrics-heavy, technical precision
   • Business (product, consulting): Business impact focus, professional warmth
   • Arts/Creative (design, media): Flexible, personality allowed, mission-driven
4. Package Outputting — Render PDFs + generate strategic interview materials:
   • Cover letter, HR/technical/user/executive interview guides, elevator pitch, outreach templates

Usage

Recommended install:

npx skills add fakhriaunur/cv-building

Works with any AI agent that supports the skills standard, including Claude Code, Claude Desktop, Cursor, Codex, Copilot, Windsurf, Gemini CLI, and more.

Manual install: Copy this directory into your agent's skills folder. Trigger with:

• "build my CV"
• "create resume"
• "prepare interview materials"
• "render CV PDF"

Agent support:

| Agent | Install | |-------|---------| | Claude Code | Copy to .claude/skills/ in your project | | Cursor | Add to .cursor/skills/ or project root | | Windsurf | Place in .windsurf/skills/ | | Codex | Reference as a skill module | | Any agent with skill support | Drop the cv-building/ folder into your skills directory |

Quick Scenarios

| Scenario | What to say | What happens | |----------|------------|--------------| | First time, no materials | "Help me build my CV from scratch" | Guided career discovery → master.md → CV → PDFs | | Have an existing CV | "Here's my current CV, build me a better one" | Extracts all data into master → quality-gated mold → PDF | | Just add a new role | "I started a new job at Company X as Senior Engineer in March 2024" | Appends to master, logs in changelog | | Target a specific role | "Build my CV for a Staff ML Engineer position at Stripe" | Reads master → STEM-gated mold → cv_stripe_staff-ml-engineer.yaml → PDF | | Need interview prep | "Prepare me for a technical interview at a fintech company" | Generates technical-interview.md + deep-dive.md from master | | Need a cover letter | "Write a cover letter for this job posting: [paste]" | Generates cover-letter.md tailored to the posting | | Use a different identity | "Create a separate master for my consulting work as 'Alex ML'" | Creates master_consulting.md as a separate persona | | Just render | "Render my existing CV with the moderncv theme" | Reads cv.yaml → renders PDF with selected theme |

Prerequisites

uv tool install "rendercv[full]"
# Or fallback to pip:
# pip install "rendercv[full]"

Optional:

# Quick PDF page counting: qpdf --show-npages <file.pdf>

Design Philosophy

No footer. Generated CVs default to show_footer: false. Footers waste space on 1-page CVs, repeat info from the header, and make documents look like generated reports. Clean edges = professional polish.

Output

| File | Location | |------|----------| | master.md | cv_files/master.md (shared across all targets) | | CHANGELOG.md | cv_files/CHANGELOG.md (tracks all changes) | | CV mold (YAML) | cv_files/cv_<company>_<position>.yaml | | PDFs | Root dir: cv_<company>_<position>_classic.pdf, etc. | | Strategic materials | cv_files/strategic-package_<company>_<position>/ |

Example: cv_files/cv_stripe_senior-backend.yaml → ./cv_stripe_senior-backend_classic.pdf

Comparison

| Dimension | cv-building (this skill) | fractal-cv-builder (OG) | rendercv-skill (official) | |-----------|-------------------------|------------------------|---------------------------| | Core focus | Career pipeline: accumulate → quality gate → render + strategic materials | Achievement discovery + CV rendering | RenderCV mechanics: schema, design, CLI | | Input model | Zero-assumption, fractal accumulation from any source | Structured discovery interview | Direct CV YAML editing | | Quality gating | Industry-flexible (STEM strict, Business moderate, Arts flexible) | Anti-cringe protocol (STEM only) | Schema validation only | | Strategic materials | Cover letter, HR/technical/user/executive interview guides, elevator pitch, outreach | Strategic package (elevator pitch, interview opener) | None — rendering only | | RenderCV engine | Aligned to official rendercv-skill for correctness | Custom theme bundles (now outdated) | Full engine authority | | Master doc | Premix storage — neutral, infinite potential, changelog-tracked | Achievements inventory | None | | Multi-mold | Yes — suffixed cv_<company>_<position>.yaml for each target | Single CV at a time | Single YAML at a time | | Multi-persona | master_<persona>.md for stage names, anon identities | Not supported | Not applicable |

Credits

• fractal-cv-builder — the original skill that proved engineers don't lack achievements, they lack a systematic way to remember and articulate them. Inspired the accumulation concept.
• rendercv-skill — the official RenderCV AI skill. Authoritative source for schema, design patterns, CLI usage, and engine mechanics. This skill's RenderCV integration is aligned to rendercv-skill for correctness.

License

MIT