Litero AI — agentic threat model
Litero AI is a low-risk academic writing assistant with limited autonomy, primarily posing data privacy and intellectual property risks through PDF ingestion and web-based research capabilities.
OWASP AIVSS score rationale
| Autonomy of Action | 0.20 | |
| Goal-Driven Planning | 0.30 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.40 | |
| Persistent Memory | 0.20 | |
| Contextual Awareness | 0.40 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.50 | |
| 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 third-party commercial LLMs for text generation and paraphrasing. Key threats include prompt injection to bypass safety filters or generate academic misconduct material, and model misalignment leading to hallucinated citations despite claims of accuracy.
Processes user-uploaded PDFs ('ask your PDF') and retrieves data from academic libraries and online sources. This introduces risks of malicious PDF uploads exploiting parser vulnerabilities, data exfiltration of unpublished research, and indirect prompt injection via poisoned academic sources or web search results.
Not certain from the listing — likely employs a proprietary RAG and orchestration framework to coordinate PDF querying and web search. Threats include insecure tool integration, such as Server-Side Request Forgery (SSRF) during online research, and prompt injection manipulating the citation generation logic.
Not certain from the listing — deployed as a closed-source SaaS platform. Primary threats include inadequate sandboxing of the PDF parsing environment, exposing the host to remote code execution, and insecure storage of user-uploaded manuscripts.
Not certain from the listing — no explicit mention of security guardrails, logging, or drift monitoring. The 'AI detector & humanizer' feature suggests a focus on output styling rather than safety monitoring, leaving potential blind spots for abusive or malicious prompt patterns.
Not certain from the listing — likely relies on standard SaaS user authentication. Compliance risks center on intellectual property ownership of AI-generated academic work, plagiarism detection evasion, and GDPR/CCPA compliance regarding uploaded research data.
Operates as a standalone, single-agent writing assistant with no multi-agent or marketplace interactions described. Ecosystem threats are minimal, restricted to potential future integrations with third-party reference managers or academic databases.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).