cyber-resume-reviewer-skill README

The project's README from github.com/mubix/cyber-resume-reviewer-skill. As one markdown file: README.md.

IT and Cybersecurity Resume Reviewer

An Agent Skill for evidence-led resume reviews, job-description tailoring, exact edits, diagnostic scoring, and complete rewrites across IT and cybersecurity roles.

A full review produces an editable Markdown report and a styled PDF generated from the same source. The skill does not rank candidates, predict interviews, or invent missing achievements.

Prompt it

Attach the original resume when possible. A PDF or DOCX lets the agent check extraction and visible layout; pasted text supports content review only.

With a job description:

Use the cyber-resume-reviewer skill to review my attached resume against the attached job description. Give me a candid fit assessment, prioritized findings, and exact edits supported by my resume. Do not invent metrics or experience. Deliver the full report as Markdown and PDF.

Without a job description:

Use the cyber-resume-reviewer skill to review my attached resume for IT and cybersecurity roles. Identify the strongest evidence, the highest-value fixes, and one or two plausible target directions. Mark layout checks you cannot perform as not assessed. Deliver the full report as Markdown and PDF.

For a rewrite:

Use the cyber-resume-reviewer skill to rewrite my resume for a security engineering role. Use only facts in the source. Put questions for stronger claims outside the clean resume.

Download and install

Tagged releases publish two clean archives on the Releases page:

Neither archive contains repository documentation, tests, workflow files, the changelog, or .gitignore.

Host Install and invoke Support notes
Codex CLI or IDE Extract the .skill contents to ~/.agents/skills/cyber-resume-reviewer/, then use $cyber-resume-reviewer or let Codex select it. Codex supports local Agent Skills. Standalone and plugin distribution differ by surface.
Claude Code Extract to ~/.claude/skills/cyber-resume-reviewer/ for personal use or .claude/skills/cyber-resume-reviewer/ in a project. Invoke /cyber-resume-reviewer. Claude Code follows the Agent Skills format and supports bundled files and scripts. See Claude Code skills.
Claude web Upload cyber-resume-reviewer-claude.zip under Customize > Skills. Custom skills require Skills and code execution to be enabled. Account and workspace controls apply. See Claude custom skills.
ChatGPT In the desktop app, open Skills in the sidebar and select an installed skill with @. OpenAI documents standalone skills for ChatGPT desktop, but its public guide does not promise direct .skill upload. ChatGPT web and mobile require plugin distribution, which this repository does not provide. See OpenAI skill docs.
Gemini CLI Run gemini skills install ./cyber-resume-reviewer.skill, then /skills reload. Gemini CLI accepts local .skill packages. See Gemini CLI skill management.
Gemini Spark Copy or rename the .skill file to .zip, then upload it from Spark's Skills page. Spark accepts a ZIP with SKILL.md at its root. Region, subscription, and account limits apply. See Gemini Apps skills.

The core format follows the Agent Skills specification. Install paths, upload rules, script permissions, and feature availability remain host-specific.

What the review covers

The skill uses four lenses:

Lens Intent
Machine-read Check text extraction, reading order, and parser risks without claiming an ATS outcome.
Human-skim Test the first impression, hierarchy, clarity, and positioning.
Human-believe Check whether scope, ownership, technical claims, and results have evidence.
Human-act Align the document with a stated target while preserving requirement wording and unknowns.

A standard report selects from these sections:

Section Intent
Assessment State the target, main conclusion, strongest evidence, and highest-impact repair.
Prioritized findings and exact edits Connect each observation to source text and give a supported action or replacement.
Target fit Separate employer requirements from candidate evidence and reviewer inference.
First impression and format Report checks performed on the source file and mark unavailable checks unassessed.
Diagnostic score Provide an optional editorial score with the selected profile and assessed coverage.
Next actions Order the smallest set of changes that improves the submission.

Package map

PDF requirements

The review instructions work without the bundled renderer. Matching the supplied PDF design requires code execution plus:

Install the optional Python set with:

python3 -m pip install -r cyber-resume-reviewer/scripts/report-requirements.txt

The bundled rendering path does not require network access. It writes page images beside the PDF so the agent can inspect every page before delivery.

Releases

Pushing a v* tag runs .github/workflows/publish-skill.yml. The workflow builds both archives from a runtime allowlist, records SHA-256 checksums, and attaches the files to the matching GitHub Release. A manual run builds the same files as a workflow artifact for testing.

cyber-resume-reviewer-skill (mubix/cyber-resume-reviewer-skill)

Searching Database…

184 ·GitHub
Cyber-resume-reviewer-skill
What is this?This is an Agent Skill that reviews and rewrites IT and cybersecurity resumes using evidence from the source document, producing a Markdown report and a matching styled PDF.
Is this a standalone app or a skill I load into an AI agent?It is a Skill package (SKILL.md plus reference files, templates and scripts) that you install into a coding agent or chat assistant like Claude Code, Codex, ChatGPT, or Gemini, not a standalone program.
How do I install and invoke it?Download the tagged .skill or Claude zip archive from Releases, extract it into the host's skills folder (e.g. ~/.claude/skills/cyber-resume-reviewer/ for Claude Code or ~/.agents/skills/ for Codex), then invoke it like /cyber-resume-reviewer or by letting the agent select it.
Do I need anything beyond the host agent to get the PDF output?The review itself works with just the agent, but matching the bundled PDF design needs code execution plus Python 3.9+, PyYAML, beautifulsoup4, Python Markdown or pandoc, and WeasyPrint or wkhtmltopdf, plus pdftoppm/pdftotext/pdfinfo/pdffonts for visual checks.
Will it exaggerate my experience or guarantee interviews?No — it explicitly refuses to rank candidates, predict interview outcomes, or invent achievements, and will mark layout checks as unassessed when it cannot run them.
Which agent platforms actually support it right now?Codex CLI/IDE, Claude Code, Claude web (via zip upload), ChatGPT desktop, and Gemini CLI/Spark are listed, though ChatGPT web/mobile and some feature support vary by host.
Does the PDF rendering need internet access?No, the bundled rendering path works without network access and writes page images beside the PDF for inspection.
What is the best way to prompt it for a job-specific review?Attach the original resume (PDF or DOCX works best for layout checks) along with a job description and ask it to give a candid fit assessment, prioritized findings, and exact edits without inventing metrics.
What's inside the standard report?It can include an Assessment, Prioritized findings and exact edits, Target fit, First impression and format, an optional Diagnostic score, and Next actions, built from four lenses: machine-read, human-skim, human-believe, and human-act.
Can I get a complete rewrite instead of just feedback?Yes, you can ask it to rewrite your resume for a specific role using only facts from the source, flagging anything that would need stronger evidence outside the clean version.
How are releases built and verified?A GitHub Actions workflow triggered by a version tag builds both release archives from a runtime allowlist, records SHA-256 checksums, and attaches them to the matching GitHub Release.
What are similar projects?ASu-skills, ai-aug-edu, software-certificate-skill, ASu-resume-skills, qiaomu-anything-to-notebooklm, rendercv, jinba-consulting-pptx-skill, handout-remake
Which topics is it in?resume, claude-code-skill, hacking, document-processing, career, workflow, cybersecurity, skill, pdf, markdown, agent-skills, productivity
Creator
mubix
Size
0.1 MB
Stars
184
Updated
CommitSep 20Releasev4.1Sep 07
Age
1mo
Commits · 90 Days