ad-creative
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes external performance data (CSV, API output, etc.) to optimize ad copy, which introduces a surface for indirect prompt injection.
- Ingestion points: Performance data provided by the user or fetched via platform-specific reporting tools as described in
SKILL.md. - Boundary markers: There are no specific instructions or delimiters mentioned to isolate ingested data from the agent's core instructions.
- Capability inventory: The skill relies on local CLI tools (
google-ads.js,linkedin-ads.js,tiktok-ads.js) that can perform network operations and campaign modifications. - Sanitization: No documentation is provided regarding the sanitization or validation of data retrieved from these external sources.
- [COMMAND_EXECUTION]: The instructions describe using local CLI tools for retrieving reports and managing ad campaigns, such as
node tools/clis/google-ads.js. These commands are part of the intended functional workflow. - [EXTERNAL_DOWNLOADS]: The
references/generative-tools.mdfile suggests the use of third-party tools that require downloading content from external repositories, such as cloningjamiepine/voiceboxfrom GitHub and runningnpxcommands forremotion. These are provided as setup instructions for the user.
Audit Metadata