ads
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by design.
- Ingestion points: Instructions in SKILL.md direct the agent to read context from files like .agents/product-marketing.md, .claude/product-marketing.md, and product-marketing-context.md. It also processes landing page URLs provided by the user.
- Boundary markers: There are no explicit delimiters or warnings to ignore instructions embedded within the ingested context files.
- Capability inventory: The skill is designed to manage advertising campaigns, budgets, and targeting across multiple platforms (Google Ads, Meta, LinkedIn, TikTok), utilizing tool integrations such as the Google Ads MCP.
- Sanitization: The instructions do not specify any validation or sanitization of the content retrieved from external marketing files.
- [SAFE]: The skill provides standard implementation code for tracking pixels from established advertising platforms.
- Evidence: References to domains such as googletagmanager.com, facebook.net, licdn.com, and tiktok.com in references/conversion-tracking.md are for legitimate services from well-known vendors.
Audit Metadata