ApplyIQ — agentic threat model

Not certain from the listing — likely relies on third-party LLMs for resume parsing and matching. A key threat is indirect prompt injection, where a malicious job description contains hidden instructions that trick the model into exfiltrating the user's resume or PII.

Not certain from the listing — processes and stores highly sensitive user resumes (PII) and job preferences. Threats include unauthorized access to the underlying database/vector store and data leakage during the matching process.

Not certain from the listing — orchestrates the workflow of matching resumes to Adzuna's database and executing applications. Threats include insecure tool integration where the agent is manipulated into submitting applications to malicious phishing sites.

Not certain from the listing — hosted on Adzuna's infrastructure. A major threat is Remote Code Execution (RCE) during the parsing of untrusted user-uploaded resume files (PDF/Docx) if the parsing environment is not properly sandboxed.

Not certain from the listing — claims 'no bias' and 'quality over quantity' but provides no details on guardrails. Gaps in observability could allow the agent to silently fail or submit incorrect data to employers without the user's knowledge.

Not certain from the listing — handling job applications and resumes requires strict alignment with GDPR, CCPA, and equal opportunity employment regulations. The listing does not specify OAuth controls or credential management for external job boards.

Not certain from the listing — integrates with Adzuna's wider job search network and external ATS (Applicant Tracking Systems). Threats include cascading failures or data exposure if partner APIs or external job boards are compromised.