agentleFS
jpitajpita/job-hunt-kit-template/skills/job-fit/SKILL.md

job-fit

Skill75 starsChanged 56 days agoMITView on GitHub

Judge whether a job posting is worth applying to, then build the tailored CV only after the user says go. Use this whenever the user pastes a job description, a job link (LinkedIn, Wellfound, Otta, Ashby, Greenhouse, Lever, a careers page), a recruiter message, or a referral link, even with no instruction. Also use it for "is this one worth it", "should I apply", "check this company", "make a CV for this", or "write a cover letter for this". A pasted posting with no instruction means: research the company, score the fit, give a verdict, and wait. Do not start a CV before the verdict is on screen.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. ZoolCoder/scoutjobs★ 4same name
    2. lengrufeng2008/dsh-skill-library★ 0same 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 repositoryOne other file it gives its agents.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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
  1. Distinctive:

    If the fetch has no requirements section (login wall, empty shell), say so and ask for the text.

    line 30
  2. Very specific:

    Search job-log.md for the company.

    line 42
  3. Very specific:

    Read references/company-research.md.

    line 51
  4. Very specific:

    Read work-history.txt now.

    line 55
  5. Distinctive:

    Never edit a focus CV to fit one posting.

    line 108
  6. Very specific:

    Wait for OK. Then follow references/cv-build.md.

    line 118
  7. Very specific:

    Read me/preferences.md and me/work-history.txt fresh, every time.

    line 122
  8. Distinctive:

    Nothing goes public: No posts, no public repos naming a company the user applies to.

    line 133

≈ “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.

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

Job fit

A pasted posting is a question: "is this worth my time". Answer it before any CV work.

Find the kit

The kit root is two folders above this file. Resolve the symlink:

KIT="$(cd "$(dirname "$(readlink -f ~/.claude/skills/job-fit/SKILL.md)")/../.." && pwd)"

Personal files, all in $KIT/me/:

  • preferences.md: hard filter, location, timezone, roles wanted.
  • work-history.txt: the source of truth for what the user did.
  • profile.py, cv-data/*.py: CV facts and focus CVs.
  • job-log.md: every past verdict.

Re-read each file in the same turn you use it. Never score from memory.

Half one: decide

1. Get the posting text

  • Pasted text is the posting.
  • For a link, load WebFetch via ToolSearch and fetch it.
  • If the fetch has no requirements section (login wall, empty shell), say so and ask for the text. Never guess.

Pull out, and say which the ad does not state:

  • employment type and duration
  • remote, hybrid or on-site, and which countries or time zones
  • hours and required overlap
  • level and years asked for
  • pay and currency
  • the interview process

2. Check the log

Search job-log.md for the company. If found, open the report with "You looked at this on <date> and decided <verdict>".

3. Hard filter

Apply the hard filter from preferences.md. If it fails, stop. Quote the line of the ad. Do no research. Everything else is a flag, not a stop.

4. Research the company

Read references/company-research.md. Every company claim carries its source link in the same line.

5. Score the fit

Read work-history.txt now. Put each requirement in one bucket:

CallMeans
StrongA named job did this, and work-history.txt says so.
PartialReal but smaller: less time, a side of another job, or a personal project.
GapNot done, nothing close.
Cannot claimA banned tool, or marked NOT confirmed in work-history.txt.
  • A keyword match is never a claim. An interviewer's follow-up on a tool never used costs more than the missing keyword.
  • Count years honestly. A job that started 8 months ago is 8 months, not "2+ years".

6. Verdict

Lead with one of:

  • GO: clears the real requirements, company checks out.
  • GO WITH CAVEATS: name the one problem in the same sentence.
  • SKIP: name the one deciding reason.

Pick a side. The table carries the nuance.

Report format

## <Company> - <Role>
Remote <yes/no> · <type, duration> · <countries> · <hours + overlap> · Comp: <found, or "not stated">

**<VERDICT>** - one sentence why.

**Company**
- 3 to 5 lines, each with its source link.

**Fit**
| Ask | Evidence | Call |
|---|---|---|

**Cannot claim** - tools and claims to keep off the CV.

Then append one line to job-log.md (newest first) and stop. No CV yet.

Half two: after the user says go

Check the focus CVs first

List me/cv-data/*.py. Read the closest one or two. Ask: would a recruiter find the ad's asks in the first 15 lines (role, tagline, top 3 skills rows)? If yes, name that PDF in me/cvs/ and stop. This is the common case.

Make a new CV only when:

  • the level differs from every focus CV
  • the ad's core sits across two or more focus CVs
  • a real must-have is missing from the closest CV
  • the domain matters and no focus CV speaks to it

Never edit a focus CV to fit one posting.

If a new one is needed, show the plan first

  • which focus file gets copied, and why none fits
  • the ROLE and TAGLINE
  • the summary angle in one sentence
  • the SKILLS row order
  • which gaps get named

Wait for OK. Then follow references/cv-build.md.

What this kit is tuned for

Read me/preferences.md and me/work-history.txt fresh, every time. This kit has no fixed role, level or country of its own: whatever those two files say at the moment you score a posting is the bar. A US-only or UK-only "remote" role is a flag when it excludes a place in me/preferences.md: check if they hire contractors there before ruling it out.

Things that go wrong

  • Leading senior on a role a level down. Reads as overqualified. Match the pitch to the ad.
  • An intermediary is not the employer. Who signs the contract, who you report to, whose product you build: answer all three.
  • The wish list is not the bar. A missing nice-to-have is not a skip. A missing core must-have is.
  • The process is a decision input. Rounds, take-home length: put them in the header.
  • Nothing goes public. No posts, no public repos naming a company the user applies to.

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.