git:20260723.de898c0 to git:20260906.85504a9

4 added, 4 removed. Audit A to A.

---

name: ats-analyzer

description: Score CV-to-job-offer match, identify missing keywords, and check ATS parseability. Use before submitting any application.

- metadata: {"navin":{"emoji":"🧮","category":"careers"}}

+ metadata: {"navin":{"emoji":"🧮","category":"careers","default_for":"career"}}

---

# ATS Analyzer

## Overview

Simulate what an ATS and a keyword-scanning recruiter see: match score, gaps, and parsing risks.

## Analysis dimensions

### 1. Keyword match

- Extract from the offer: hard skills, tools, certifications, methodologies, soft skills, degree requirements

- Classify each: ✅ present in CV / 🟡 present but different wording / ❌ absent

- Score = weighted coverage (must-haves count triple)

### 2. Wording alignment

Flag synonym mismatches the ATS may miss: "gestion de projet"≠"project management", "JS"≠"JavaScript", acronyms without expansion (write both: "Search Engine Optimization (SEO)").

### 3. Parseability

- Single column? standard section headers? no text boxes/images for text?

- Dates in consistent format? contact info in body (not header/footer)?

- File type: DOCX or text-layer PDF (test: can you select the text?)

## Output format

```markdown

- ## ATS analysis — <CV> vs <offer>

+ ## ATS analysis - <CV> vs <offer>

Match score: X% (must-haves: Y/Z)

|-------------|--------|--------|-----|

### Parsing risks

### Top 5 actions before applying

```

## Workflow

- 1. Ingest offer + CV (`pdf-ocr-extractor` for PDFs).

+ 1. Ingest the Career opportunity + Master CV (`career action=status` / import). `pdf-ocr-extractor` only for a user-provided PDF.

2. Build the requirement table; score honestly.

3. For each ❌: is it hidden in the candidate's real experience? → surface it (via `cv-tailoring`). Truly missing? → say so; suggest addressing in the cover letter or skipping the application.

## Rules

- The score guides effort; below ~50% must-have coverage, recommend not applying.

- - Never advise keyword-stuffing invisible text or lying — both backfire.

+ - Never advise keyword-stuffing invisible text or lying - both backfire.