ae-drama-creative-evaluation
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill employs bash and Python scripting within its workflow to dynamically render SQL templates and process JSON-formatted metadata retrieved from the analytics platform.
- [COMMAND_EXECUTION]: Legitimate command execution occurs through local scripts (
precheck.sh,metadata_scan.sh) that utilize theae-clitool to verify project authorization and scan available event metadata. - [INDIRECT_PROMPT_INJECTION]: The skill processes structured data from external SQL queries and metadata exports. The potential for indirect prompt injection is mitigated by a mandatory 'Confirmation Gate' in Step 4, requiring the user to manually verify field mappings and data feasibility before the analysis proceeds.
- [SAFE]: Analysis of the instructions and scripts revealed no evidence of obfuscation, hardcoded credentials, unauthorized network activity, or attempts to bypass security guidelines.
Audit Metadata