visual-qa-testing
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run
npm run devto start a local development server, which is a routine operation in software development workflows. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection because it reads and processes data from external web pages, console logs, and network requests.
- Ingestion points: Web content is ingested via
browser_navigate, while logs and network data are read viabrowser_console_messagesandbrowser_network_requestsin SKILL.md. - Boundary markers: Absent; the instructions do not include delimiters or warnings to ignore malicious instructions embedded in the browser output.
- Capability inventory: The skill allows the agent to execute shell commands (
npm run dev) and interact with page elements (browser_click,browser_fill), which could be abused if an injection occurs. - Sanitization: Absent; there is no evidence of validation or filtering for the data ingested from the browser tools.
Audit Metadata