laravel-vite
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to analyze user-controlled configuration files, creating a vector for indirect instruction injection.\n
- Ingestion points: Mandatory workflow in
SKILL.mdrequires checkingvite.config.jsandpackage.jsonthrough thefuse-ai-pilot:explore-codebaseagent.\n - Boundary markers: No specific delimiters or safety warnings for processing external file content are provided in the workflow instructions.\n
- Capability inventory: The agent has the capability to execute build commands (
npm run build,vite) and modify the filesystem based on findings from the codebase exploration.\n - Sanitization: No validation, filtering, or sanitization of configuration file content is mentioned before the agent uses it for implementation tasks.
Audit Metadata