frontend-performance
Fail
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: HIGHPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill instructs the user to execute 'sudo safaridriver --enable' in 'SKILL.md' to configure the Safari WebDriver on macOS, which requires administrative privileges.
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to download and install development tools and packages from trusted sources, including 'npx playwright install webkit' (from Microsoft) and '@next/bundle-analyzer' (from Vercel).
- [COMMAND_EXECUTION]: The skill provides instructions for running various shell commands such as 'open -a Safari ', 'next build', and 'safaridriver' for testing and build verification.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external performance auditing tools like Lighthouse and PageSpeed Insights, which could potentially contain untrusted instructions.
- Ingestion points: Performance audit reports and failing audits from Lighthouse and PageSpeed Insights.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious embedded content within the audit reports.
- Capability inventory: The skill allows for shell command execution ('open', 'next build', 'safaridriver') and uses package managers ('npm', 'npx').
- Sanitization: There is no evidence of sanitization or strict validation of the audit tool outputs before they are processed by the agent.
Recommendations
- AI detected serious security threats
Audit Metadata