agentleFS
squernesquerne/open-career-skills/CLAUDE.md

open-career-skills

CLAUDE.md26 starsChanged 3 months agoMITView on GitHub

This is a career workspace, not a codebase. The user forked it to run elite career-asset skills locally with their own Claude subscription. Your job across every task here: produce career assets grounded ONLY in the user's real facts. Never invent a fact about the user. No numbers, no titles, no achievements, no enthusiasm they didn't express. When a fact is missing, ask, or leave a visible [add ...] placeholder. A polished fabrication is the worst output this workspace can…

The file

We split this CLAUDE.md into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow each skill's process exactly, especially the hard stops: where a skill says to ask the user and wait, end your turn.

    line 43

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

Open Career Skills workspace

This is a career workspace, not a codebase. The user forked it to run elite career-asset skills locally with their own Claude subscription. Your job across every task here: produce career assets grounded ONLY in the user's real facts.

The one law of this workspace

Never invent a fact about the user. No numbers, no titles, no achievements, no enthusiasm they didn't express. When a fact is missing, ask, or leave a visible [add ...] placeholder. A polished fabrication is the worst output this workspace can produce; recruiters reject it and interviews expose it.

Corollary (groundedness): every factual claim in a generated asset must trace to a line in the profile, the story bank, or the user's own answers. The /apply pipeline enforces this with an independent reviewer; hold yourself to the same bar everywhere else.

Layout and state

  • profile/profile.md: the user's career profile, created by /setup from profile/profile.template.md. Read it before any career task. If it doesn't exist, suggest /setup first.
  • story-bank/: one STAR story per file, named YYYY-MM-short-slug.md (format in story-bank/README.md). This is the user's achievement memory; skills read it for evidence and write to it via extraction.
  • content/calendar.md: LinkedIn pillars and post log, created from content/calendar.template.md on first use.
  • tracker/applications.md: the application log, created from tracker/applications.template.md by /apply. The memory of the pipeline: dedup for /find-jobs, rehearsal targets for the mock interviewer, gap history for /upskill.
  • output/: generated CVs, cover letters, interview debriefs, per-application folders (output/apply-<company>/).
  • documents/: (optional, user-created) source materials like an exported CV.
  • docs/: WORKFLOW.md (the pipeline), SKILLS.md (per-skill contracts), TROUBLESHOOTING.md.

Privacy invariant: .gitignore keeps profile/profile.md, personal story-bank/*.md files, content/calendar.md, output/, and documents/ out of git so a public fork can never leak the user's CV, stories, or metrics. Never weaken those rules, never write personal data into a *.template.md file, and never stage an ignored file with git add -f.

Output discipline: deliverables (CVs, letters, stories, debriefs) are written to their file first; the terminal gets a brief summary, the file path, and the skill's closing line, not a 500-word wall. Each skill specifies this; follow it.

Output style: never write an em dash character in a generated asset (CV, letter, story, post). It is the single most recognized AI tell in application documents; recruiters and ATS-adjacent tools flag it. Use a comma, colon, or plain hyphen as the sentence demands. The application-reviewer enforces this; don't override that critique.

Routing

User intentUse
Tailor/rewrite CV for a jobruthless-cv-optimizer skill (or /optimize-cv)
Capture an achievement, prep behavioral stories, match stories to a JDstar-story-extractor skill (or /extract-story)
LinkedIn ideas or post draftslinkedin-360brew-planner skill (or /plan-content)
Cover lettercover-letter-writer skill (or /cover-letter)
Interview practicemock-interviewer skill (or /mock-interview)
Find current job postingsjob-search skill (or /find-jobs)
Apply to a specific job end-to-end/apply <url or JD> (fit gate, drafts, reviewer, tracker)
Mine git history for achievementsevidence-miner skill (or /mine-evidence)
Which ATS a company uses, and the CV rules it impliesats-detective skill (or /detect-ats)
Recurring skill gaps / what to learn/upskill
Build/refresh the profile/setup

Follow each skill's process exactly, especially the hard stops: where a skill says to ask the user and wait, end your turn. Do not answer your own questions or skip a phase to be helpful.

Privacy note to keep in mind

Everything the user stores here stays on their machine; the .gitignore privacy guard keeps personal files out of commits even on public forks. If the user asks to commit or share one of those files, warn them once about what becomes public, then follow their call.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.