AR

artbyjazi/resume-strategist

Developer tools
11 stars Quality 57 Trend 16MIT

Open-source resume strategist skill for students. Tailors resumes + cover letters in any LLM. Evidence-based. MIT.

Overview

An open-source, model-agnostic "resume strategist" skill for students and early-career candidates. It does what an $800–1,000 resume-writing package does — strategy conversation, tailored resume, cover letter, quality review — as a single instruction file any LLM can follow. Every rule is tied to recruiter eye-tracking studies, ATS recruiter interviews, or 2025–26 hiring surveys (see references/evidence.md). — Settings → Skills → upload SKILL.md. It triggers automatically when you share a resume or job posting. — clone straight into your skills directory: — open PROMPT.md, copy all of it, paste it as your first message (or into custom instructions / a custom GPT / a Gem). Then paste your resume and the job posting. 1. Intake — works from a resume, a posting, or both 2. Strategy conversation — 3–5 pointed questions, not a questionnaire 3. Job-posting decode — must-haves, exact vocabulary, knockout filters 4.

README

resume-strategist

An open-source, model-agnostic “resume strategist” skill for students and early-career candidates. It does what an $800–1,000 resume-writing package does — strategy conversation, tailored resume, cover letter, quality review — as a single instruction file any LLM can follow. Every rule is tied to recruiter eye-tracking studies, ATS recruiter interviews, or 2025–26 hiring surveys (see references/evidence.md).

Quickstart

Claude (claude.ai / desktop app) — Settings → Skills → upload SKILL.md. It triggers automatically when you share a resume or job posting.

Claude Code — clone straight into your skills directory:

git clone https://github.com/artbyjazi/resume-strategist.git ~/.claude/skills/resume-strategist

ChatGPT, Gemini, or any other chat — open PROMPT.md, copy all of it, paste it as your first message (or into custom instructions / a custom GPT / a Gem). Then paste your resume and the job posting.

Ollama / local models — use PROMPT.md as the system prompt.

What it does

  1. Intake — works from a resume, a posting, or both
  2. Strategy conversation — 3–5 pointed questions, not a questionnaire
  3. Job-posting decode — must-haves, exact vocabulary, knockout filters
  4. Tailored one-page resume — built for the 7-second F-pattern scan, parse-clean for ATS search
  5. Cover letter — paragraph or skills-match-table format, cliché-banned
  6. Anti-AI-tell pass + quality gate — six-axis score and three next actions

What it refuses to do

Invent metrics, titles, tools, or experience. It asks for numbers instead of making them up, and won’t pad a skills section the bullets don’t back.

Repo structure

SKILL.md                   the skill — frontmatter + instructions
PROMPT.md                  generated from SKILL.md, frontmatter stripped, paste anywhere
references/evidence.md     the sources behind each rule
scripts/sync_prompt.py     regenerates PROMPT.md; --check fails on drift
scripts/check_frontmatter.py  validates the SKILL.md frontmatter

PROMPT.md is generated — edit SKILL.md and run python scripts/sync_prompt.py. CI enforces it.

Roadmap

  • Role vocabulary packs (SWE, finance, consulting, marketing)
  • Country-specific norms (Canadian co-op vs US vs UK/EU)
  • Before/after example resumes
  • A companion edition for mid-level and executive candidates

Contributing

PRs welcome — see CONTRIBUTING.md. The short version: never hand-edit PROMPT.md, every rule needs a source in references/evidence.md, and no individual coaches or services get named.

License

MIT — see LICENSE.

View the full README on GitHub

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add artbyjazi/resume-strategist
New to MCP? Start with the mcpskill.org guide →