Koke AI — agentic threat model
Koke AI is a low-risk, single-purpose citation utility with minimal agentic capabilities, presenting low overall security risk primarily limited to prompt injection and potential SSRF if resolving user-provided URLs.
OWASP AIVSS score rationale
| Autonomy of Action | 0.10 | |
| Goal-Driven Planning | 0.00 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.10 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.20 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.30 | |
| Opacity & Reflexivity | 0.20 |
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 relies on a third-party foundation model (e.g., GPT-4o or Claude) to parse text and format citations. Primary threats include prompt injection to bypass formatting rules or generate inappropriate content.
Not certain from the listing — may ingest user-provided text, PDFs, or URLs to extract metadata. Risks include data leakage of sensitive student research papers and potential Server-Side Request Forgery (SSRF) if fetching external URLs.
Not certain from the listing — likely uses a simple, non-agentic LLM chain rather than a complex agent framework. Risks are limited to basic prompt manipulation affecting the output format.
Not certain from the listing — hosted as a standard web application. Standard web security risks (XSS, CSRF, insecure API endpoints) apply, particularly if user inputs are rendered directly.
Not certain from the listing — likely lacks advanced LLM-specific observability or guardrails, relying instead on basic input validation and standard web logging.
Not certain from the listing — as a freemium educational tool, it likely lacks formal enterprise compliance certifications (e.g., SOC2, FERPA) and robust data retention policies.
Not certain from the listing — operates as a standalone vertical application with no multi-agent interactions or ecosystem dependencies described.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).