show-me-your-work
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes task-related data to generate a decision log. It mitigates potential injection risks by utilizing the
scripts/log.shutility, which sanitizes inputs by removing control characters (tabs, newlines) and prefixing spreadsheet formula triggers (symbols like =, +, -, and @) with a single quote. - [COMMAND_EXECUTION]: A local utility script
scripts/log.shis used for file operations. The script employs defensive programming techniques, including strict shell options and consistent variable quoting, ensuring safe execution within the agent's environment.
Audit Metadata