ScreenSmart — agentic threat model
ScreenSmart presents moderate agentic risk, primarily driven by the handling of sensitive candidate PII and the potential for prompt injection attacks embedded in resumes to manipulate hiring scores. Its lack of autonomous real-world action limits systemic damage, but compliance and data privacy risks remain high.
OWASP AIVSS score rationale
| Autonomy of Action | 0.30 | |
| Goal-Driven Planning | 0.20 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.10 | |
| Persistent Memory | 0.20 | |
| Contextual Awareness | 0.40 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| 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 — likely uses commercial LLMs to parse and evaluate resumes. Primary threats include indirect prompt injection (e.g., candidates hiding white-text instructions in resumes to force a 100/100 score) and inherent algorithmic bias in candidate evaluation.
Not certain from the listing — processes highly sensitive candidate PII (contact info, work history). Threats include unauthorized data retention, lack of encryption at rest/in transit, and potential data leakage if candidate resumes are inadvertently used for model training.
Not certain from the listing — relies on orchestration to ingest resumes, compare them to job descriptions, and output scores. Threats include insecure document parsing libraries (vulnerable to PDF/Docx exploits) and manipulation of the scoring logic via malformed inputs.
Not certain from the listing — deployed as a paid API. Threats include weak API authentication, lack of multi-tenant isolation allowing recruiters to see other companies' candidate data, and insecure container hosting.
Not certain from the listing — requires continuous monitoring to detect drift, bias, and adversarial resume formats. Without explicit observability, systematic discrimination or automated screening bypasses could go unnoticed.
Not certain from the listing — must align with strict regulations governing automated employment decision tools (AEDTs) such as NYC Local Law 144 and GDPR Article 22. The listing does not mention independent bias audits or data deletion policies.
The listing describes ScreenSmart as a standalone vertical API tool with no multi-agent coordination or marketplace integrations, minimizing ecosystem-specific cascading risks.
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.