ads

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides automated workflows in references/creative-research-automation.md for scraping external data from the Facebook Ad Library and product review sites (Amazon, G2, Trustpilot). This processed data could contain malicious instructions. However, the skill includes explicit security guardrails in references/audit-guardrails.md and references/creative-research-automation.md that instruct the agent to treat external content strictly as data and to ignore any embedded directives like "ignore previous instructions."
  • [COMMAND_EXECUTION]: The skill interacts with advertising platforms (Google Ads, Meta, LinkedIn) through integrated tools. It enforces a strict "draft first" policy in references/audit-guardrails.md, requiring the agent to propose changes (current state, proposed change, and rollback steps) for explicit user approval before execution.
  • [EXTERNAL_DOWNLOADS]: The skill references and provides installation instructions for official tracking pixels and conversion APIs from well-known services including Google, Meta, LinkedIn, and TikTok in references/conversion-tracking.md. These are standard marketing industry practices and target official vendor domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 11:19 AM