web-tooling-vite
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [SAFE]: Analysis of the instructions and code examples shows they represent standard development workflows for Vite configuration, with no execution of suspicious commands or unauthorized file access.
- [PROMPT_INJECTION]: The skill describes patterns for reading and modifying project configuration files (e.g.,
vite.config.ts,tsconfig.json). While these represent a surface for indirect prompt injection from codebase content, the skill follows standard assistant patterns and provides security guidance, such as warning against committing secrets in.envfiles. - [DATA_EXFILTRATION]: The configuration examples utilize
loadEnvto access environment variables. This is a standard feature of the tool being configured, and the skill correctly notes the importance of ignoring secret-bearing files like.env.localto prevent data exposure in build artifacts or source control.
Audit Metadata