building-ui-bundle-frontend
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection due to its core function of generating and building code from user instructions.
- Ingestion points: User requests for UI modifications, including new pages and components, serve as the entry point for potentially malicious instructions (SKILL.md, implementation/page.md).
- Boundary markers: The skill does not provide the agent with specific delimiters or instructions to isolate or sanitize user-provided implementation details.
- Capability inventory: The skill allows for writing and modifying files across the project (src/pages/, src/components/, etc.) and directs the agent to execute shell commands for linting and building the application within the local environment (SKILL.md, implementation/page.md).
- Sanitization: There are no instructions for validating or escaping user-provided data before it is incorporated into the application code or processed by build tools.
Audit Metadata