browser-app-delivery
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The instructions require the agent to execute shell commands for environment discovery, such as probing for executables using
command -v chromium, starting application servers, and performing network reachability checks. - [INDIRECT_PROMPT_INJECTION]: The skill derives build and execution parameters from local project files and manifests, creating an attack surface where malicious project metadata could influence agent behavior.
- Ingestion points: Local project structures, application manifests, and build configuration files.
- Boundary markers: The instructions include strict constraints on when to load scaffolding and how to format the final delivery report.
- Capability inventory: Shell execution, server process management, and binding to network interfaces for accessibility.
- Sanitization: The skill does not define specific validation or sanitization logic for strings extracted from project files before they are used in commands.
- [DYNAMIC_EXECUTION]: The verification workflow involves executing JavaScript artifacts within a temporary harness to validate initialization and render cycles before the project is handed off to the user.
Audit Metadata