deliver
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill explicitly requires the agent to 'run actual commands' to verify task completion. Examples include
npx tsc --noEmitand running tests. While intended for validation, this provides a functional mandate for broad shell access. - [PROMPT_INJECTION]: The skill creates an 'Indirect Prompt Injection' surface by processing user-provided requests into 'targets' and 'done-conditions'. These generated targets then drive the automated command execution and verification loop without explicit sanitization or boundary markers between the user input and the system commands.
Audit Metadata