webapp-testing
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The helper script
scripts/with_server.pyusessubprocess.Popenwithshell=Trueto launch local servers defined by the user or agent. This is a standard pattern in development runners to support shell features like directory changes (cd) and command chaining (&&) required by most web framework startup commands (e.g.,npm run dev). - [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to automate web browsers to interact with applications, creating a surface for indirect prompt injection if the content being tested contains instructions targeted at the agent.
- Ingestion points:
page.content(),page.locator().all(), and visual analysis of browser screenshots. - Boundary markers: The instructions do not define specific delimiters or warnings to separate processed web data from the agent's internal logic.
- Capability inventory: The skill allows for local file writing (automation scripts), execution of these scripts, and management of local server processes via
subprocesscalls inscripts/with_server.py. - Sanitization: No explicit sanitization or filtering of external web content is described in the provided instructions.
Audit Metadata