---
name: skills-assessment
description: Design, deploy, and grade candidate skills assessments. Use when you need to evaluate technical abilities, soft skills, role-specific competencies, or cultural alignment before interviews. Supports coding challenges, case studies, personality inventories, timed tests, proctored exams, and automated grading. Triggers on phrases like "assess candidates", "create a skills test", "coding challenge", "case study evaluation", "evaluate technical skills", "assessment workflow", "pre-screen test".
---

# Skills Assessment

Evaluate candidate competencies at scale before advancing to interview stages.

## Workflow

1. Analyze the job description to identify which skills must be validated (technical, functional, soft skills).
2. Select or generate the appropriate assessment type per skill category (see assessment matrix).
3. Configure assessment parameters: duration, passing threshold, proctoring level, accommodation options.
4. Send invitations with clear instructions, deadlines, and tech-requirements checklist.
5. Monitor completion rates; send reminders at 24h and 48h for incomplete assessments.
6. Auto-grade objective assessments; flag subjective assessments for reviewer assignment.
7. Generate scorecards per candidate with breakdown by competency area.
8. Cross-reference assessment scores with resume screening scores; compute combined ranking.
9. Advance candidates who meet threshold; draft personalized feedback for those who don't.

## Assessment Type Matrix

| Skill Category | Assessment Type | Duration | Auto-Gradeable | Tools |
|---------------|----------------|----------|----------------|-------|
| Coding / Dev | Timed coding challenge | 60–90 min | Yes | HackerRank, Codility, CodeSignal |
| Data / Analytics | SQL + visualization task | 45–60 min | Partial | DataLemur, Analytic Skills |
| Design | Portfolio review + whiteboard exercise | 90–120 min | No | Figma, Marksheet |
| Writing / Content | Timed writing sample | 30–45 min | Partial | TestGorilla, custom |
| Sales | Role-play call or email thread simulation | 20–30 min | No | RoleRaiser, custom |
| PM / Strategy | Product case study | 60–90 min | No | ProductPlan, custom |
| Soft Skills / Culture | Situational judgment test | 20–30 min | Yes | Predictive Index, HireSkill |
| Language | Fluency test (written + spoken) | 30–45 min | Partial | LangTest, custom |

## Assessment Design Principles

### Avoid Unnecessary Barriers

- **Time limit**: Set duration at 1.5× what a qualified candidate needs. If median completion is 40 min, set limit to 60 min.
- **No trick questions**: Assess real job skills, not puzzle-solving ability.
- **Take-home vs. timed**: Take-home for depth (48h window), timed for speed-of-thought. Use both for senior roles.
- **Accommodations**: Offer +50% time for documented disabilities, non-native language speakers, or candidates in significantly different time zones.

### Coding Challenge Guidelines

```
Structure:
  - 1 easy problem (warm-up, tests basic competency)
  - 1 medium problem (core job skill)
  - 1 hard problem (differentiates strong candidates)
  - Optional: 1 collaborative/debugging task (tests teamwork)

Evaluation criteria:
  - Correctness (does it pass test cases?) — 40%
  - Code quality (readability, structure, naming) — 25%
  - Efficiency (time/space complexity) — 20%
  - Edge case handling — 15%

Anti-gaming measures:
  - Proctoring: webcam + browser lock for high-stakes roles
  - Plagiarism detection: compare against public solutions (LeetCode, GitHub)
  - Follow-up: ask candidate to explain their solution in screening call
```

### Case Study Template (PM / Strategy Roles)

```
Prompt:
"You are the PM for [product feature]. The VP has asked you to
evaluate whether to invest in [initiative]. You have 90 minutes."

Deliverable:
  - 1-page recommendation (Go / No-Go / Pivot)
  - Market sizing estimate
  - Key risks and mitigations
  - Proposed metrics for success
  - 30-60-90 day plan if approved

Grading rubric:
  - Structure and clarity of thinking — 30%
  - Data-driven reasoning (not just opinion) — 30%
  - Awareness of trade-offs — 20%
  - Actionable next steps — 20%
```

## Scoring & Thresholds

### Per-Assessment Scoring

| Score Band | Label | Action |
|-----------|-------|--------|
| 90–100% | Exceptional | Fast-track to final round |
| 75–89% | Strong | Advance to interview |
| 60–74% | Qualified | Advance if interview slot available |
| 40–59% | Borderline | Advance only with strong resume score (> 4.0) |
| < 40% | Not ready | Reject with encouraging feedback |

### Combined Ranking Formula

```
Final Rank Score = (Resume Score × 0.40) + (Assessment Score × 0.60)

Rationale: Assessment weights higher because it validates actual ability,
not just claimed experience. However, resume score still matters for
context (years of experience, scope, progression).
```

## Non-Completion Handling

Candidates who don't complete assessments fall into buckets:

| Reason | Action |
|--------|--------|
| No response to invite | Send 1 reminder; if still no response after 72h, auto-reject |
| Started but didn't finish | Offer 24h extension once; if still incomplete, evaluate on resume alone with -0.5 penalty |
| Technical issues reported | Reschedule with 48h window; log incident for QA |
| Explicitly opted out | Respect decision; evaluate on resume alone or withdraw |

## Scorecard Output

```
Candidate Assessment Scorecard
==============================
Role: Senior Frontend Engineer
Assessment: Coding Challenge (90 min) + SJT (30 min)

Coding Challenge:
  Problem 1 (Easy — Array Manipulation):     PASS  ✓
  Problem 2 (Medium — React Component):      PASS  ✓  (minor style note)
  Problem 3 (Hard — State Management):       PARTIAL ✓ (correct logic, verbose code)
  Score: 78/100

Situational Judgment Test:
  Collaboration scenarios:    4/5
  Technical decision-making:  5/5
  Score: 90/100

Combined Assessment Score: 84/100 → STRONG
Recommendation: Advance to Technical Interview
```

## Bias Mitigation

- **Blind grading**: Hide candidate name/photo during assessment review
- **Standardized rubrics**: Every candidate graded against the same criteria — no curve grading
- **Diverse reviewer pool**: At least 2 reviewers per subjective assessment, different backgrounds
- **Audit trail**: Log all scores, reviewer IDs, and timestamps for compliance
- **Regular calibration**: Monthly reviewer calibration sessions to align scoring

## Integration Points

- Assessment platforms: HackerRank, Codility, TestGorilla, Marksheet, RoleRaiser
- ATS: Auto-import scores back into candidate record
- Proctoring: Examity, ProctorU (for high-stakes roles)
- Email: Invitation, reminder, and result notifications
- Analytics: Track assessment-to-hire correlation to validate assessment quality

## Edge Cases

- **Senior/Staff roles**: Skip generic coding challenges; use system design interview or architecture review instead
- **Bootcamp grads**: Weight practical assessment heavier than education credentials
- **International candidates**: Account for time zone in deadline setting; offer rescheduling flexibility
- **Accessibility**: Provide screen-reader-compatible assessments, alternative formats for motor impairments