claude-career-workspace
Evidence-based job-search workspace for Claude Code: resume tailoring, LinkedIn audits, outreach and interview prep that never invent experience or metrics
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/mohammed-shanid--claude-career-workspace)HTML
<a href="/listings/mohammed-shanid--claude-career-workspace"><img src="/api/listings/mohammed-shanid--claude-career-workspace/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
1. **Use this template, then make your copy private.** Your copy will hold your phone number, salary expectations and application history. 2. Put your resume in `reference/resume/` and your LinkedIn PDF export in `reference/linkedin/`. 3. Open the folder in Claude Code and run:
/start-here # collects your evidence and resolves conflicting facts /linkedin-audit # section-by-section critique of your current profile /optimize-linkedin # rewritten profile copy (only after the audit)
4. Paste a job description into a chat, or save it in `reference/job-descriptions/`, and run `/resume-optimizer`.
LaTeX resume compilation needs `pdflatex` installed locally. Everything else is plain text.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 103 lines
- # Claude Career Workspace
- **A job-search workspace for [Claude Code](https://claude.com/claude-code) that refuses to make things up.**
- Most AI resume tools will happily turn "helped with deployments" into "led a cloud migration that cut costs 40%." That gets you through a keyword filter and then sinks you in the first technical interview.
- This workspace works the other way. Every claim about you is logged in a **verification register** with its source and status. Every resume line, LinkedIn section and interview story is built only from claims that pass. When a number doesn't exist, it asks you instead of inventing one.
- It's plain Markdown, CSV and LaTeX. There is nothing to install, no database and no web app. Open the folder in Claude Code and run a slash command.
- ---
- ## What it does
- | Command | What you get |
- |---|---|
- | `/start-here` | Inventories your resume, LinkedIn export and documents, finds claims that contradict each other, and asks you to resolve them before anything gets written |
- | `/linkedin-audit` | A recruiter-lens review of all 12 profile sections, a resume-vs-LinkedIn consistency check, and 2–3 positioning options with trade-offs |
- | `/optimize-linkedin [section]` | Copy-paste-ready profile sections: 3 scored headline variations, About, Experience, Skills, Featured |
- | `/resume-optimizer <job description>` | A requirement → evidence matrix (Strong / Partial / Gap), rewritten summary and bullets with before → after → why, and a tailored one-page LaTeX resume, compiled and checked for a clean text layer |
- | `/recruiter-outreach` | Under-100-word messages and follow-ups for a named person and a real opening, logged in a tracker. Never sent for you |
- | `/interview-coach` | A STAR story bank built only from verified experience, likely questions for a specific interview, and mock interviews with feedback |
- | `/job-search-planner` | A 30-day plan sized to the hours you actually have, plus weekly reviews driven by your tracker numbers |
- | `/linkedin-content` | Technical post drafts grounded in work you really did, or feedback on your own draft |
- | `/salary-negotiator` | Negotiation prep from a real offer and cited market data only. No invented benchmarks |
- ## How the evidence system works
- ```text
- your resume, LinkedIn export, context/verification-register.md
- certificates, public repos, ──► every claim + source + status
- your answers in conversation (VERIFIED, CONFIRMED BY USER, SELF-REPORTED,
- NEEDS VERIFICATION, CONFLICTING, REJECTED)
- │
- ▼
- only VERIFIED / CONFIRMED BY USER claims
- reach approved public material
- ```
- - **Your resume and LinkedIn are treated as claims, not evidence.** When they disagree (dates, titles, client counts), the conflict is flagged and stays out of final copy until you resolve it.
- - **Verbs match the evidence.** "Studied", "used", "built" and "maintained in production" are different claims, and the skills map records which one applies.
- - **"I" and "we" stay separate.** Team results aren't attributed to you alone.
- - **Public checks are logged.** When a claim can be checked on a public page (a GitHub repo, a publication record), the URL and access date go into the register.
- - **No guarantees.** It never promises ATS selection, recruiter replies or offers, and it doesn't claim to know LinkedIn's ranking algorithm.
- ## What it will never do
- - Post, send, apply, connect or edit anything on LinkedIn, email or job portals. Everything is a draft you act on yourself.
- - Put your phone number, address, salary or anything you mark private into public-facing material.
- - Overwrite an approved document. New versions get a new date, and every substantive change is logged in `CHANGELOG.md`.
- ## Quick start
- 1. **Use this template, then make your copy private.** Your copy will hold your phone number, salary expectations and application history.
- 2. Put your resume in `reference/resume/` and your LinkedIn PDF export in `reference/linkedin/`.
- 3. Open the folder in Claude Code and run:
- ```text
- /start-here # collects your evidence and resolves conflicting facts
- /linkedin-audit # section-by-section critique of your current profile
- /optimize-linkedin # rewritten profile copy (only after the audit)
- ```
- 4. Paste a job description into a chat, or save it in `reference/job-descriptions/`, and run `/resume-optimizer`.
- LaTeX resume compilation needs `pdflatex` installed locally. Everything else is plain text.
- ## Folder map
- ```text
- CLAUDE.md Rules Claude follows in every session
- CHANGELOG.md Record of every substantive change
- context/ What's known about you, and how sure it is
- verification-register.md Every claim, with its source and status (the single source of truth)
- career-profile.md Identity, education, credentials, seniority
- experience.md Each role, with discrepancies and open questions
- skills-and-projects.md Skills mapped to evidence levels; projects
- target-roles.md Target roles grouped into tracks; chosen positioning
- preferences.md Location, work mode, constraints (some fields private)
- reference/ Your source material. Claude never edits it
- analysis/ Audits, positioning options, onboarding status
- outputs/ Drafts and approved deliverables, one folder per module
- trackers/ applications.csv, networking.csv, progress.csv
- .claude/skills/ One folder per slash command, with the full workflow
- ```
- ## Who it's for
- It was built for an early-career technical job search: cloud, DevOps, Linux, support and security roles. Some defaults reflect that, such as the interview-coach technical topics and the example job boards. The evidence rules, trackers and workflow apply to any field. Change `context/target-roles.md` and the skills follow it.
- ## Customizing
- - **Tracks and roles:** edit `context/target-roles.md` and the `track` values in `trackers/README.md`.
- - **Rules:** `CLAUDE.md` is short on purpose. Add your own guardrails, such as words to avoid or claims never to make.
- - **Workflows:** each skill is a single `SKILL.md` in `.claude/skills/<name>/`. Edit the procedure directly.
- ## Credits
- The module structure (30-day plan, resume rewrite, headline writer, cold outreach, interview stories, salary negotiation, job-search map) was inspired by a set of seven career prompts shared on social media; the original author is unknown. The evidence register, verification rules and skills in this repo are new work built on top of that idea.
- ## License
- MIT. See `LICENSE`.
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.