ad-creative
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to analyze performance data from external sources (CSV, text, or API) to identify winning patterns and generate new variations. This creates an indirect prompt injection surface where untrusted data could contain instructions meant to subvert the agent's behavior.
- Ingestion points: Processes performance data provided as CSV files, text pastes, or API outputs as described in SKILL.md.
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are mentioned in the data processing workflow.
- Capability inventory: The skill identifies capabilities for executing local CLI scripts (node tools/clis/google-ads.js) and interacting with external AI generation APIs.
- Sanitization: No specific validation, filtering, or escaping logic is defined for the incoming performance data.
- [EXTERNAL_DOWNLOADS]: The documentation provides instructions for cloning an external repository from GitHub (https://github.com/jamiepine/voicebox.git) to set up voice generation tools.
- [REMOTE_CODE_EXECUTION]: The skill recommends the use of npx create-video@latest, which involves downloading and executing code from the NPM registry at runtime to initialize video project templates.
- [COMMAND_EXECUTION]: The skill includes examples and instructions for executing local Node.js scripts (e.g., tools/clis/google-ads.js) to fetch data from and manage campaigns on external advertising platforms.
Audit Metadata