evaluate-candidate

name	evaluate-candidate
description	Evaluate a candidate's Scoutica Skill Card against a job description or role requirements. Use when asked to match a candidate to a role, score a profile, check if an opportunity fits, or pre-screen applicants. Returns a structured fit report with scores and recommendations.

evaluate-candidate

Score a candidate's Skill Card against specific job requirements and produce a structured evaluation.

When to Use

• Employer asks "does this candidate fit our role?"
• Candidate asks "do I match this job?"
• ATS system needs automated pre-screening
• Agent needs to compare multiple candidates

Step-by-Step Process

Step 1: Load the Skill Card

Read all 4 files from the candidate's card folder or URL:

profile.json  → skills, experience, seniority
rules.yaml    → engagement rules, salary floor, remote policy
evidence.json → proof of work
SKILL.md      → entry point

Step 2: Load Job Requirements

Accept requirements as either:

• A job description (parse it)
• A structured object: {required_skills, nice_to_have, seniority, salary_budget, remote_policy, industry}

Step 3: Score Skills Match

The CLI is the canonical scorer — always prefer it over hand-rolling the math:

scoutica evaluate <card> --role <role.json> [--json]

This runs tools/scoring.py, a deterministic engine. The candidate's skill set is the union of profile.skills, tools_and_platforms, specializations, and primary_domains. The formula is:

hard_ratio      = matched(role.requirements.hard_skills)      / count(role.requirements.hard_skills)
preferred_ratio = matched(role.requirements.preferred_skills) / count(role.requirements.preferred_skills)

skills_score = hard_ratio × 70 + preferred_ratio × 30

# Bonuses (the ONLY two; final score is capped at 100):
+10  if evidence.json demonstrates ≥ 50% of the hard skills
+5   if candidate seniority exactly matches role.requirements.seniority

Step 4: Check Rules of Engagement

# From rules.yaml — check each rule:
salary_check:     job.salary >= rules.compensation.minimum_base_eur
remote_check:     job.remote_policy ∈ rules.remote.policy
industry_check:   job.industry ∉ rules.filters.blocked_industries
engagement_check: job.engagement_type ∈ rules.engagement.allowed_types

If ANY hard filter fails → HARD_REJECT (do not continue).

Step 5: Verify Evidence (Optional)

For each claimed skill in evidence.json:

• Check if URL is reachable
• Verify GitHub repos match claimed languages
• Check certification validity dates

Step 6: Generate Fit Report

{
  "candidate": "Name",
  "role": "Senior DevOps Engineer",
  "overall_score": 78,
  "verdict": "STRONG_MATCH",
  "skills_analysis": {
    "matched": ["Kubernetes", "Terraform", "AWS", "Python"],
    "missing": ["Ansible"],
    "bonus": ["Go", "Helm"],
    "skills_score": 80,
    "bonus_score": 10
  },
  "rules_check": {
    "salary" 
     
     
     
  
   
   
   

Scoring Thresholds

Score	Verdict	Action
80-100	STRONG_MATCH	Proceed to interview
≥60	MODERATE_MATCH	Review and decide
40-59	WEAK_MATCH	Consider if other factors compensate
0-39	NO_MATCH	Do not proceed
Any hard filter fails	HARD_REJECT	Auto-reject, do not present

Privacy Rules

• Use ONLY Zone 1 (public) data for initial screening
• Zone 2 (verified) data requires authentication
• NEVER access Zone 3 (private) data without explicit candidate approval
• NEVER share evaluation results with the candidate without employer consent
• NEVER share candidate data with third parties

Anti-Discrimination

• NEVER factor in name, gender, age, ethnicity, nationality, photo
• Evaluate ONLY on: skills, experience, evidence, engagement rules
• Include this disclaimer in every report: "Evaluation based solely on verified skills and stated requirements"