Transparent scoring

Five anchored dimensions — zero black boxes

Every answer is evaluated by our 3-Pass AI engine — Evaluator, Challenger, Reconciler — against a published rubric with concrete written anchors at every score band. Candidates see exactly how they were evaluated. Hiring managers get defensible, explainable scores.

Dimension 1

Threat Identification

Identifies attack vectors, techniques, and threat actor behavior — with precise MITRE mapping at top scores.

Dimension 2

Containment & Response

Minimizes blast radius while preserving evidence — compliance-aware (CERT-In, GDPR, DPDPA).

Dimension 3

Architecture & Blast Radius

Maps architectural interconnections and impact — surfaces weaknesses and suggests segmentation.

Dimension 4

Communication Quality

Clear, structured, audience-appropriate — from board briefing to technical deep-dive.

Dimension 5

Framework Application

Correct application of MITRE ATT&CK, NIST, CIS, ISO 27001 — and knows when they don't apply.

Each question weights these dimensions differently — a "map this to MITRE" question weights Threat Identification at 40%; an "executive briefing" question weights Communication at 50%. The full weighting matrix is in our published rubric.

The 3-Pass AI Engine

Pass 1

Evaluator

Scores the answer against the anchored rubric with full per-dimension justification.

Pass 2

Challenger

Adversarially reviews the Evaluator's reasoning and challenges any unsupported scores.

Pass 3

Reconciler

Resolves disagreements with a defensible final score — the version you and the hiring manager see.

Read the full scoring rubric

Assessments are evaluated by Anthropic's Claude API across three sequential passes per question. The full anchored rubric, qualitative bands, confidence calculation, and per-category weighting are published. Transparency is the trust differentiator.