← browser-testing-with-devtools
browser-testing-with-devtools — agentic threat model
The browser-testing-with-devtools agent presents a high-risk profile due to its browser-automation capabilities via Chrome DevTools MCP. If compromised via prompt injection, it could be coerced into navigating to malicious sites, exfiltrating sensitive session data, or performing SSRF attacks.
OWASP AIVSS score rationale
| Autonomy of Action | 0.70 | |
| Goal-Driven Planning | 0.60 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.50 | |
| Dynamic Identity | 0.20 | |
| Multi-Agent Interactions | 0.40 | |
| Non-Determinism | 0.50 | |
| Opacity & Reflexivity | 0.40 |
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 — relies on an external LLM to drive the MCP tool calls, exposing it to prompt injection that could hijack browser automation.
Not certain from the listing — primarily inspects runtime DOM/network data rather than managing a vector database, but sensitive data in transit (session tokens, PII in DOM) could be exfiltrated.
Integrates with Chrome DevTools MCP server. Threat: Tool misuse (e.g., navigating to malicious sites, executing arbitrary JS via console, exfiltrating network requests).
Requires and orchestrates an external MCP server and a real browser instance. Threat: Browser-automation surface, potential container escape, local network access via the browser, or SSRF-like behavior if the browser is not sandboxed.
Not certain from the listing — lacks built-in guardrails or logging mechanisms mentioned in the description to detect malicious navigation or data exfiltration.
Not certain from the listing — no mention of authentication, authorization, or policy enforcement for restricting which domains the browser can visit.
Uses MCP (Model Context Protocol) to interact with an external server. Threat: Compromised MCP server or malicious tool definitions leading to cascading failures or unauthorized browser control.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).