lab
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes a 'Reporting' section that instructs the agent to use security-related terminology such as 'Severity' (HIGH, MEDIUM, LOW) and 'Block/Approve' status. While these are intended for quality assurance of the generated component lab, the use of these keywords mimics security audit findings.
- [COMMAND_EXECUTION]: The instructions require the agent to execute local shell commands to 'Start the dev server and open the lab'. This is standard functionality for a development-focused skill and does not involve suspicious network or file system operations.
Audit Metadata