v0-dev
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONSAFE
Full Analysis
- [EXTERNAL_DOWNLOADS] (INFO): The skill documentation (references/sharp_edges.md) encourages the use of CLI tools that fetch remote content.
- Evidence: Recommendations to use 'npx shadcn-ui@latest add' and 'npx v0 add' to install UI components and dependencies.
- [COMMAND_EXECUTION] (INFO): The reference files contain various shell commands intended for the user to manage their project environment.
- Evidence: 'ls components/ui/', 'npx shadcn-ui@latest init', and grep-based checks for hardcoded colors in component files.
- [INDIRECT_PROMPT_INJECTION] (INFO): The skill is designed to process untrusted user input (UI descriptions) to generate output, creating a potential surface for prompt injection, though the capability is limited to display.
- Ingestion points: User prompts for UI generation ('generate ui', 'generate component').
- Boundary markers: None present in the skill instructions to delimit user input from system instructions.
- Capability inventory: The skill generates code blocks and UI architectures for user review; it has no direct file-write or network capabilities in its defined logic.
- Sanitization: No explicit sanitization or validation of user-provided descriptions is performed before processing.
- [SAFE] (SAFE): No evidence of malicious obfuscation, credential theft, persistence mechanisms, or unauthorized privilege escalation was found across the analyzed files.
Audit Metadata