agentleFS
miteshjadav98miteshjadav98/ClaudeSkills/skills/job-match-scoring/SKILL.md

job-match-scoring

Skill0 starsChanged 3 months agoMITView on GitHub

LinkedIn job-scrape scoring workflow — recomputes a blended weighted match score over scraped job data (weighting differentiator skills, title-fit, and seniority), filters and ranks, and produces a ranked JSON report plus an Excel tracker. Use whenever the user wants to score, rank, filter, or triage scraped job listings, run a job search over LinkedIn, or build a job-match report or tracker.

The file

Files
SKILL.md

Job Match Scoring

Score and rank scraped job listings against the candidate's profile, then produce a ranked report and tracker. The scraper's own keyword score is a raw signal only — it counts literal string presence, so a strong-fit role loses points when the JD omits terms like "MCP" or "FastAPI", or uses synonyms ("AOAI" for Azure OpenAI). Always recompute a blended weighted score in post-processing.

Candidate profile

Fill this from the candidate before scoring: target titles, years of experience, and locations of interest (including remote). The example weighting below is tuned for an AI/GenAI Engineer track — target titles such as AI/GenAI Engineer, LLM Engineer, Conversational AI Engineer, or Backend Engineer. Retune the skill universe and weights for a different track.

Blended score

Example skill universe (18): Python, FastAPI, Node.js, TypeScript, Azure OpenAI, LangChain, LangGraph, RAG, Vector Database, LLM, Agentic AI, MongoDB, PostgreSQL, Redis, Docker, Azure, MCP, Prompt Engineering.

Differentiator weights (the candidate's high-signal skills weigh most):

SkillWeight
LangGraph3.0
RAG3.0
Agentic AI3.0
Azure OpenAI2.5
FastAPI2.5
MCP2.5
all other skills1.0

Score = (sum of weights of matched skills ÷ sum of weights of all skills) × 100, then apply:

  • Title-fit bonus for core AI roles (title contains AI / GenAI / LLM / Machine Learning / Agentic).
  • Penalty for architect / QA / support / non-IC titles that don't match an IC engineering track.
  • Optionally a small seniority-fit adjustment (senior/lead roles fit; junior roles down-weight).

Filter to ≥55%. Sort by blended score → AI relevance → recency (most recent first).

Apify scraper gotchas (actor: cheap_scraper/linkedin-job-scraper)

Enum values are hyphenated — underscores are silently ignored:

  • jobType: "full-time" (not full_time)
  • experienceLevel: "associate", "mid-senior" (not underscore forms)
  • workType: "remote", "hybrid", "on-site"

resumeKeywords takes objects with keyword + optional aliases; returns keywordMatchScorePercentage per job — treat as raw signal only. Paginate get-dataset-items with offset/limit; dedupe across pages by jobUrl.

Outputs

  1. Ranked JSON — full structured array, one object per job with all requested fields (title, company, location, workType, experienceLevel, postedTime, blendedScore, rawScore, matchedKeywords, unmatchedKeywords, salaryInfo, applicationsCount, jobUrl).
  2. Excel tracker — top N rows with matched/missing skills, salary, applicant counts, clickable apply links, plus working columns: Status, Applied Date, Notes. Read the xlsx skill first.
  3. A short in-chat summary table (rank, blended %, title, company, location) with a color cue (🟢 for ≥80%) and a one-line note on recurring resume gaps to close (which high-weight terms are frequently unmatched).

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.