Desklib — agentic threat model
Desklib exhibits a low-risk agentic profile, primarily functioning as an information retrieval, recommendation, and content generation tool for education. Its primary security risks lie in data integrity (poisoned scholarly sources) and prompt injection affecting quiz generation, rather than autonomous execution or systemic privilege escalation.
OWASP AIVSS score rationale
| Autonomy of Action | 0.20 | |
| Goal-Driven Planning | 0.20 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.20 | |
| Persistent Memory | 0.40 | |
| Contextual Awareness | 0.30 | |
| Dynamic Identity | 0.10 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.40 | |
| Opacity & Reflexivity | 0.30 |
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 utilizes commercial LLMs for quiz generation and research assistance. Vulnerable to prompt injection (e.g., jailbreaking the quiz generator) and indirect prompt injection via malicious text embedded in uploaded or indexed scholarly documents.
Not certain from the listing — relies on an integrated digital library of vetted scholarly materials and likely a vector database for RAG. Vulnerable to knowledge-base poisoning if the vetting process for scholarly materials is bypassed, and data exfiltration of proprietary academic texts.
Not certain from the listing — orchestrates recommendation engines and quiz generation. Vulnerable to insecure tool integration, particularly if the document parsers handling multi-format learning materials (PDFs, docs) are susceptible to remote code execution or denial of service.
Not certain from the listing — deployed with mobile integration and web access. Vulnerable to standard mobile OWASP Top 10 risks, insecure API endpoints, and lack of sandboxing during the processing of user-uploaded study materials.
Not certain from the listing — no public details on guardrails or hallucination monitoring. Gaps here could allow the agent to output confidently incorrect academic information or biased recommendations without detection.
Not certain from the listing — closed-source, freemium model. Requires alignment with student data privacy standards (such as COPPA or FERPA depending on the target demographic) and robust access control to protect premium/vetted content from unauthorized scraping.
Not certain from the listing — operates as a standalone vertical educational tool. There is no indication of multi-agent orchestration or third-party agent marketplace integration, minimizing ecosystem-level cascading risks.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).