Lightscreen AI — agentic threat model
Lightscreen AI presents a moderate risk profile primarily centered on candidate data privacy (PII) and the potential for prompt injection to game interview evaluations. As an AI recruiting tool, it operates in a high-scrutiny domain (employment) where bias, non-determinism, and compliance (e.g., EU AI Act) are critical concerns.
OWASP AIVSS score rationale
| Autonomy of Action | 0.40 | |
| Goal-Driven Planning | 0.30 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.20 | |
| Persistent Memory | 0.30 | |
| Contextual Awareness | 0.40 | |
| Dynamic Identity | 0.10 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.50 | |
| Opacity & Reflexivity | 0.40 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — The specific foundation models used for conversational interviewing and technical evaluation are not disclosed. Threats include prompt injection by candidates attempting to bypass technical questions or manipulate the grading criteria, as well as model reprogramming.
Not certain from the listing — The agent processes highly sensitive candidate PII, resumes, and live interview transcripts/recordings. Threats include data exfiltration of candidate records and potential knowledge-base poisoning if custom rubrics are manipulated.
Not certain from the listing — The orchestration framework managing the interview flow, question generation, and grading is unspecified. Threats include session hijacking or memory poisoning during an active interview session to alter the evaluation outcome.
Not certain from the listing — While the tool is open source, the deployment architecture (SaaS vs. self-hosted) is not detailed. Threats include unauthorized access to hosted interview sessions, API key exposure, and lack of sandboxing for any live code execution environments used during technical tests.
Not certain from the listing — Although marketed as 'uncheatable', the specific guardrails, anti-cheating mechanisms, and evaluation monitoring tools are not detailed. Threats include candidates finding bypasses to the anti-cheat logic or exploiting evaluation drift.
Not certain from the listing — No specific compliance standards (such as SOC2, GDPR, or EU AI Act high-risk AI system requirements) are mentioned. Threats include regulatory non-compliance regarding automated employment decision-making and algorithmic bias.
Not certain from the listing — The agent likely integrates with Applicant Tracking Systems (ATS) to export candidate scores. Threats include compromised ATS API integrations leading to unauthorized data modification or lateral movement into broader HR systems.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.