resume-agent
Skill181 starsChanged todayMITView on GitHub
Create, edit, validate, and render Resume-Agent JSON resumes into local XeLaTeX PDFs, or sync resumes with the Resume-Agent Web workspace via the resume-agent CLI. Use when the user asks to draft a new resume, tailor or update a ResumeData JSON file, export a resume PDF locally, or pull and push resumes to Resume-Agent.
- Installs packages
2 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- wycocoyu/Resume-Agentedited, first seen Nov 15, 2026
- couimet/my-claude-skillstheir own version, first seen Nov 1, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- wycocoyu/Resume-Agent★ 18189% same
Wrote their own version · 1
- couimet/my-claude-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- AGENTS.mdAGENTS.md
- web/AGENTS.mdAGENTS.md
- web/CLAUDE.mdCLAUDE.md
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
Keep validation read-only and mandatory: Always run resume-agent validate before render or sync push.
- Very specific: line 19
Sync only on explicit user intent: Never run resume-agent sync push or overwrite local files with --force unless the user asked for it.
- Very specific: line 51
Before editing fields, read references/schema.md for the exact ResumeData structure, field conventions, and built-in logo keys (or inspect …
- Very specific: line 65
If exit code is 1, inspect errors[*].path (for example menuSections[0].enabled or education[0].school) and repair the JSON.
- Very specific: line 92
Add --online only when the user wants remote COS/HTTP logo or avatar downloads.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Resume Agent
Build, update, validate, render, and optionally sync Workspace v2 ResumeData JSON resumes using the deterministic resume-agent CLI, while this harness handles natural-language authoring and tailoring.
Invoke this skill in Codex with $resume-agent.
Core Guardrails
- Never fabricate candidate facts. Do not invent employers, schools, degrees, dates, GPAs, project metrics, open-source links, or awards. Populate only facts stated by the user or present in user-provided source files. When key facts (such as exact metrics, project scale, performance improvements, dates, awards) are missing or not provided by the candidate, mark them explicitly with
【待补】(e.g.峰值 QPS 由 2000 提升至【待补】) or ask concise follow-up questions, or disable unused sections viamenuSections[*].enabled = false. Never invent quantitative metrics or employment histories. - Keep validation read-only and mandatory. Always run
resume-agent validatebeforerenderorsync push. Fix reported field paths directly in the JSON file. - Stay offline by default.
resume-agent renderruns with zero network access unless the user explicitly asks to fetch remote COS/HTTP logos or avatars via--online. - Sync only on explicit user intent. Never run
resume-agent sync pushor overwrite local files with--forceunless the user asked for it. - Page balance and density self-check. After rendering, check whether the resume fills the page cleanly. Avoid sparse overflow where a second page contains only 2-3 orphan lines. Tailor bullet points and margins so that a 1-page resume is compact and complete, or a 2-page resume is substantively filled on both pages.
Prerequisites & Skill Discovery
- Editable CLI install (from the Resume-Agent repository root):
pip install --no-build-isolation -e /path/to/Resume-Agent - Harness discovery for Codex (writes into the documented skills directory):
resume-agent skill install --harness codex --dry-run resume-agent skill install --harness codexresume-agent skill status --harness codexreports drift and CLI/skill version skew;resume-agent skill uninstall --harness codexremoves only files it created.
Standard Workflow
1. Initialize or Load Resume JSON
- Start a new resume:
resume-agent init -o resume.json --title "后端开发简历" --json - Pull an existing resume from Web workspace (when requested):
resume-agent sync list --json resume-agent sync pull <resume_id> -o resume.json --json
2. Author or Edit ResumeData
Before editing fields, read references/schema.md for the exact ResumeData structure, field conventions, and built-in logo keys (or inspect the live machine schema via resume-agent schema).
Critical formatting rules:
- 8 HTML rich-text fields (
education[*].description,workExperience[*].details,experience[*].details,projects[*].description,openSource[*].description,customData[*][*].description,selfEvaluation,skillContent): format using clean HTML such as<ul class="custom-list"><li><p><strong>关键词</strong>:具体职责与量化结果</p></li></ul>or<p>...</p>. - Plain-text fields (especially
awards[*].description, plus all names, titles, companies, schools, roles, and dates): write plain text ONLY without HTML tags (awards[*].descriptionis LaTeX-escaped directly, so<p>would print literally in the PDF). - Section visibility and ordering: control rendered sections via
menuSections(enabled: true/falseand array order). Anycustom_*key incustomDatamust have a matching enabled entry inmenuSections.
3. Validate Deterministically
resume-agent validate resume.json --json- Exit code
0("valid": true) is required before rendering or pushing. - If exit code is
1, inspecterrors[*].path(for examplemenuSections[0].enabledoreducation[0].school) and repair the JSON.
4. Render Local Resume (HTML, Headless Chrome CDP PDF, or XeLaTeX)
- Render self-contained HTML:
cococv render resume.json -o resume.html --template asu-tech --json - Render A4 PDF via Headless Chrome CDP (recommended for HTML templates):
cococv render resume.json -o resume.pdf --template asu-tech --json - Render XeLaTeX PDF (for LaTeX templates):
cococv render resume.json -o resume.pdf --template default --engine latex --json
Template Recommendation Matrix
asu-tech(Recommended default): High-density single-column technology template, grey capsule badges, multi-level STAR bullet points. Ideal for technical and engineering roles.asu-01~asu-04,asu-07,asu-10,asu-11,asu-15: Single-column distinctive themes (classic blue accent line, solid header band, vertical timeline, ribbon badge, serif typography).asu-05,asu-13: Two-column sidebar themes (light grey sidebar, pure black sidebar) ideal for separating skills and basic info on the left with experience on the right.asu-06,asu-08,asu-09,asu-12,asu-14,asu-16~asu-18: Two-page / high-density / academic themes (timeline two-page, cards, ultra-dense two-column, pure black academic).default: XeLaTeX classic single-column layout, optimal for academic or ATS plaintext extraction.
Notes:
- Uses local XeLaTeX or headless Chrome offline (
offline: true). - Local PNG/JPEG/PDF avatars in
basic.photo(file path ordata:image/png;base64,...) are validated by magic bytes at the CLI boundary; WebP/SVG or missing logos degrade gracefully and are reported inwarnings. - Add
--onlineonly when the user wants remote COS/HTTP logo or avatar downloads.
5. Optional Web Workspace Sync
- Push validated JSON to Web workspace:
resume-agent sync push resume.json --name "张三-后端开发" --alias "秋招投递版" --json - Authentication options:
- Bearer session token:
--token <token>orRESUME_AGENT_TOKEN(takes strict precedence and suppresses internal secret headers). - Local trusted dev headers:
FASTAPI_INTERNAL_AUTH_SECRET+RESUME_AGENT_USER_ID(or--internal-secret+--user-id). - Target server:
--api-urlorRESUME_AGENT_API_URL(supports both direct FastAPIhttp://127.0.0.1:9000and Next.js proxyhttp://127.0.0.1:3000/api/fastapi/proxy).
- Bearer session token:
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

