ad-creative
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests and processes untrusted data from the user and external context files, which represents an indirect prompt injection surface.
- Ingestion points: User-provided product information, audience data, performance metrics, and the
marketing-context.mdfile (referenced in SKILL.md). - Boundary markers: The skill does not define clear delimiters or provide instructions to the agent to disregard instructions embedded within these data sources.
- Capability inventory: The skill utilizes a local Python script (
scripts/ad_copy_validator.py) to process and evaluate generated text. - Sanitization: While the validator script checks for platform-prohibited phrases and character limits, it does not include sanitization or filtering to prevent prompt injection attacks from malicious input.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script,
scripts/ad_copy_validator.py, to validate ad copy. The script is used for legitimate text analysis tasks such as checking character counts against platform specifications and identifying common reasons for ad rejection. It does not perform network operations or access sensitive system directories.
Audit Metadata