Dessa AI — agentic threat model
Dessa AI is a low-risk, browser-based voice and text-to-speech utility with minimal agentic capabilities, presenting primary risks around prompt injection for voice generation and credit exhaustion rather than systemic infrastructure compromise.
OWASP AIVSS score rationale
| Autonomy of Action | 0.20 | |
| Goal-Driven Planning | 0.00 | |
| Self-Modification | 0.00 | |
| Dynamic Tool Use | 0.00 | |
| Persistent Memory | 0.00 | |
| Contextual Awareness | 0.00 | |
| Dynamic Identity | 0.00 | |
| Multi-Agent Interactions | 0.00 | |
| Non-Determinism | 0.70 | |
| Opacity & Reflexivity | 0.00 |
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.
Uses GPT-Live-1 and HeyGen voices. Vulnerable to prompt injection attacks that could force the model to output offensive, malicious, or brand-damaging audio content, or bypass credit limits.
Not certain from the listing — No persistent database or RAG pipeline is described. Data operations appear limited to transient processing of microphone audio, captions, and WAV generation.
No complex agent orchestration framework, memory, or tool-calling capabilities are implemented, minimizing the risk of framework-level exploits or unauthorized tool execution.
Not certain from the listing — Implemented as a browser studio. Primary infrastructure risks involve the secure handling of API keys for GPT-Live-1 and HeyGen, and potential client-side vulnerabilities like Cross-Site Scripting (XSS).
Not certain from the listing — No details are provided regarding input/output guardrails, abuse monitoring, or logging of voice interactions to prevent malicious generation.
Not certain from the listing — No compliance certifications or specific identity/access management controls are detailed, beyond the use of paid credits for generation.
Operates as an isolated, standalone browser tool with no multi-agent coordination, external marketplace integrations, or agent-to-agent trust boundaries.
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 — every score is re-derived by the same automated method as an agent's public evidence changes.