Zuni — agentic threat model
Zuni presents a moderate-to-high risk profile primarily due to its deep integration into the user's browser and Gmail context, making it highly susceptible to indirect prompt injection from untrusted web pages that could lead to data exfiltration or unauthorized email drafting.
OWASP AIVSS score rationale
| Autonomy of Action | 0.40 | |
| Goal-Driven Planning | 0.30 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.50 | |
| Persistent Memory | 0.20 | |
| Contextual Awareness | 0.80 | |
| Dynamic Identity | 0.30 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.60 | |
| Opacity & Reflexivity | 0.50 |
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 multiple frontier models (OpenAI, Anthropic, Google, DeepSeek, Mistral, Meta). Main threats include prompt injection (especially indirect injection via open tabs) and misaligned outputs during email drafting.
Ingests active browser tab contents and Gmail context. This creates a significant threat of data exfiltration if a malicious website poisons the tab context to extract sensitive Gmail data or session tokens.
Not certain from the listing — the orchestration framework for the Chrome extension is unspecified. Potential threats include insecure DOM manipulation and unsafe handling of API responses within the sidebar.
Not certain from the listing — details on how API keys for the various frontier models are secured (client-side vs. proxy backend) are not provided, posing risks of credential theft if stored insecurely in the extension.
Not certain from the listing — there is no mention of guardrails, input filtering, or logging mechanisms to detect and prevent malicious prompt injections from active web pages.
Not certain from the listing — no compliance certifications (such as SOC2) or explicit data privacy policies regarding the processing of Gmail and tab data are detailed.
Not certain from the listing — the agent appears to operate as a single-user productivity tool without explicit multi-agent coordination or marketplace integrations.
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. Are you the vendor? Factual corrections are free.