CV Validation Skill

When to Apply

Trigger this skill when:

•User asks to validate CV JSON
•Before PDF generation (always validate first)
•After manual edits to CV data
•When checking ATS compliance
•User uploads CV and asks "is this valid"
•Debugging layout overflow issues

Core Validation Workflow

Step 1: Load Schema Definition

First, understand the complete schema by reading: references/schema-requirements.md

Key schema points:

•Required fields: firstName, lastName, email, phone, address, professionalTitle
•Optional but recommended: photo_url, languages, skills, experience, education
•Size constraints: photo_url ≤32KB, bullets ≤90 chars

Step 2: Pre-Validation (Fast Local Check)

Before calling API, run deterministic validation script:

bash

python .claude/skills/cv-validation/scripts/validate_schema.py <cv-json-file>

Script checks:

•JSON syntax validity
•Required field presence
•Field type correctness (string, array, object)
•Size constraints (photo_url, bullet lengths)
•Character encoding (UTF-8, no invalid chars)

Output:

json

{
  "valid": true,
  "errors": [],
  "warnings": ["photo_url is 28KB, close to 32KB limit"]
}

Step 3: API Validation (Comprehensive)

Call Azure Function for full business rule validation:

bash

curl -X POST http://localhost:7071/api/validate-cv \
  -H "Content-Type: application/json" \
  -d @<cv-json-file>

API validates:

•Schema compliance (all pre-validation checks)
•Business rules (e.g., dates in chronological order)
•Language-specific constraints (EN/DE/PL character sets)
•2-page layout estimation

Responses:

•✅ Valid: {"valid": true, "message": "CV is valid", "layout_estimate": "1.8 pages"}
•❌ Invalid: {"valid": false, "errors": [{"field": "...", "message": "..."}]}

Step 4: Layout Space Estimation

Use script to estimate if content fits 2-page template:

bash

python .claude/skills/cv-validation/scripts/count_template_space.py <cv-json-file> --language=en

Output:

code

Estimated pages: 1.9 / 2.0
Margin: 10% (SAFE)

Breakdown:
- Experience section: 1.2 pages
- Education section: 0.3 pages
- Skills section: 0.2 pages
- Languages section: 0.1 pages
- Header/footer: 0.1 pages

Thresholds:

•≤1.8 pages: SAFE (plenty of margin)
•1.8-2.0 pages: WARNING (tight fit, test visually)
•

2.0 pages: ERROR (content overflow, must reduce)

Step 5: Report Validation Results

Format (if valid):

code

✅ CV Validation: PASS

Schema: ✅ All required fields present
Size: ✅ photo_url 28KB / 32KB, longest bullet 87 / 90 chars
Layout: ✅ Estimated 1.9 / 2.0 pages (10% margin)
API: ✅ Business rules satisfied

Ready for PDF generation.

Proceed? (yes/no)

Format (if invalid):

code

❌ CV Validation: FAIL

Errors (3):
1. [HIGH] experience[0].responsibilities[2]: Bullet exceeds 90 chars (current: 112)
   Current: "Led cross-functional team of 5 engineers to successfully deliver microservices architecture migration completing project 2 weeks ahead of schedule"
   Fix: "Led team of 5 engineers to deliver microservices migration, completing 2 weeks early" (87 chars)

2. [HIGH] photo_url: Size exceeds 32KB limit (current: 45KB)
   Fix: Compress image or reduce resolution (target: <30KB for safety margin)

3. [MEDIUM] experience[0].startDate: Date format invalid (current: "2020-Jan")
   Fix: Use ISO format: "2020-01-01"

Layout: ⚠️ Estimated 2.3 pages (exceeds 2-page limit)

Cannot proceed to PDF generation until errors are fixed.

Edit CV data? (yes/no)

Common Validation Errors

Required Field Missing

Error: Missing required field 'email' Fix: Add email to CV JSON

json

{
  "email": "john.doe@example.com"
}

Photo URL Too Large

Error: photo_url exceeds 32KB limit (current: 45KB) Root cause: Azure Table Storage property size limit Fix:

•Compress image with lossless compression
•Reduce image dimensions (200x200px sufficient)
•Convert to WebP format (better compression)
•Store in blob storage and use reference URL instead

Script to compress:

bash

python .claude/skills/cv-validation/scripts/compress_photo.py <image-file> --max-size=30KB

Experience Bullet Too Long

Error: experience[0].responsibilities[1] exceeds 90 chars (current: 112) Root cause: Template has limited space per bullet Fix:

•Remove filler words ("successfully", "effectively", "efficiently")
•Use abbreviations (e.g., "implemented" → "built", "collaborated with" → "worked with")
•Split into two bullets if genuinely complex
•Focus on impact, remove process details

Before:

code

"Successfully collaborated with cross-functional stakeholders to effectively implement enterprise-wide authentication system serving over 10,000 users across multiple regions"

After:

code

"Built enterprise auth system for 10K+ users across multiple regions" (68 chars)

Content Overflow (>2 Pages)

Error: Content exceeds 2-page limit (estimated: 2.3 pages) Root cause: Too much experience, too many skills, or verbose bullets Fix:

•Limit experience to last 10 years or 5 most relevant roles
•Reduce bullets per role (max 4-5 per position)
•Consolidate similar skills
•Remove outdated skills/certifications
•Shorten professional summary

Priority for reduction:

•Oldest experience entries (>10 years ago)
•Less relevant roles (if applying to specific position)
•Redundant skills (e.g., "JavaScript", "ES6", "React" → "React (JavaScript/ES6)")
•Long bullets (reduce to 60-70 chars if possible)

Invalid Date Format

Error: experience[0].startDate: Date format invalid (current: "2020-Jan") Root cause: Schema requires ISO 8601 format Fix: Use YYYY-MM-DD format

json

{
  "startDate": "2020-01-01",
  "endDate": "2022-12-31"
}

Handling "Present":

json

{
  "startDate": "2020-01-01",
  "endDate": null  // or omit field entirely
}

Validation Scripts

validate_schema.py

Location: scripts/validate_schema.py Purpose: Fast local JSON schema validation Usage:

bash

python .claude/skills/cv-validation/scripts/validate_schema.py <cv-json>
python .claude/skills/cv-validation/scripts/validate_schema.py <cv-json> --strict

Returns: JSON with validation results Execution time: ~50ms (fast pre-check)

count_template_space.py

Location: scripts/count_template_space.py Purpose: Estimate if content fits 2-page template Usage:

bash

python .claude/skills/cv-validation/scripts/count_template_space.py <cv-json> --language=en

Algorithm:

•Counts characters per section
•Applies language-specific multipliers (German words longer than English)
•Estimates lines based on template CSS
•Accounts for margins, headers, footers

Returns: Page estimation with safety margin

compress_photo.py

Location: scripts/compress_photo.py Purpose: Compress photo to meet 32KB limit Usage:

bash

python .claude/skills/cv-validation/scripts/compress_photo.py <image-file> --max-size=30KB --output=<output-file>

Strategies:

•Lossless compression (PNG optimization)
•Format conversion (PNG → WebP)
•Dimension reduction (maintain aspect ratio)
•Quality reduction (JPEG quality 80-90)

Progressive Disclosure

Level 1 (Metadata): Skill name + description (always loaded) Level 2 (This file): SKILL.md body (loaded when skill triggers) Level 3 (References):

•Read references/schema-requirements.md for complete schema
•Read references/ats-compliance.md for ATS-specific rules
•Read references/layout-constraints.md for 2-page template details

Only load references when:

•User asks for "detailed schema" or "complete reference"
•Validation errors are unclear and need deeper explanation
•Debugging complex layout issues

ATS Compliance Checks

For strict ATS compliance (when --strict flag used):

Additional validations:

•No tables in experience section (ATS parsers fail on tables)
•No images except profile photo
•No colored text (some ATS strip formatting)
•Standard section headers (Experience, Education, Skills)
•Phone number in standard format (E.164 or national)
•Email address valid format
•No special characters in name fields
•PDF/A compliance (archival format)

See: references/ats-compliance.md

Integration with Other Commands

Typical workflow:

•User uploads CV
•/validate-cv → Pre-check + API validation
•If valid: Generate preview HTML
•/visual-regression → Screenshot + compare baseline
•If visual OK: Generate PDF
•If invalid: Show errors, ask for fixes, return to step 2

Chaining:

code

User: "Validate this CV and generate PDF if valid"

Claude:
1. /validate-cv data.json
2. [If valid] Call generate-cv-action API
3. [If invalid] Show errors, abort PDF generation

Performance Notes

Validation timing:

•Local schema check: ~50ms
•API validation: ~200ms
•Layout estimation: ~100ms
•Total: <400ms (fast feedback loop)

Optimization:

•Cache schema definition (don't re-read on every validation)
•Parallelize local + API validation
•Skip layout estimation if pre-validation fails

Error Recovery

If validation API fails:

•Fall back to local validation only
•Warn user: "API unavailable, basic validation only"
•Skip business rules (dates, language constraints)
•Proceed with PDF generation at user's risk

If script execution fails:

•Check Python availability: python --version
•Install dependencies: pip install -r requirements.txt
•Fall back to manual validation (read schema, check fields)

References

Read these files for detailed information:

•references/schema-requirements.md - Complete JSON schema
•references/ats-compliance.md - ATS-specific validation rules
•references/layout-constraints.md - 2-page template constraints

Project files:

•Schema source: ../../DATA_DICTIONARY.md
•API implementation: ../../src/schema_validator.py
•Template spec: ../../templates/html/CV_template_2pages_2025.spec.md

Last updated: 2026-01-22 Skill version: 1.0.0