vite
Warn
Audited by Gen Agent Trust Hub on Aug 5, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill metadata identifies 'Anthony Fu' as the author, which contradicts the actual developer 'julianromli' provided in the author context. This mismatch in attribution can be misleading and may lead users or agents to accord the skill an unearned level of trust based on the reputation of the purported author.- [DATA_EXFILTRATION]: Configuration examples in 'references/core-config.md' and 'references/build-and-ssr.md' promote the use of 'loadEnv' with an empty prefix string. This pattern disables the default security filtering in Vite, causing all environment variables (including sensitive credentials like 'DB_PASSWORD' or secret tokens) to be loaded into the build context, which increases the risk of data exposure.- [EXTERNAL_DOWNLOADS]: The skill documentation includes commands to install and use various Node.js packages, such as 'vite@8', 'rolldown-vite', and various Vite plugins. It also references external documentation and resources from domains like 'vite.dev' and 'nitro.build'. These are documented as part of the tool's standard ecosystem.
Audit Metadata