creative-director

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions and reference materials are professional in nature and focused entirely on the domain of advertising and creative direction. No evidence of prompt injection, data exfiltration, or persistence mechanisms was found.
  • [METADATA_POISONING]: A discrepancy exists between the author name provided by the platform (immamdouhaboammar) and the author cited in the internal markdown documentation (Serge Shima). This appears to be a case of redistribution of an open-source tool rather than a malicious attempt to deceive, as all author-related URLs point to legitimate professional sites.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting user-supplied creative briefs. While this provides a surface for indirect injection, the skill mitigates this by requiring the agent to follow a strict 'Phase Router' and 'Creative Constitution'. These components force the AI to evaluate data against specific strategic constraints and 'Simplicity as Violence' principles, effectively acting as internal boundary markers that isolate untrusted input from the agent's core logic.
  • [EXTERNAL_DOWNLOADS]: The legendary-campaigns library contains hundreds of URLs to reputable advertising awards sites (D&AD, Cannes Lions), industry news (Fast Company, Adweek), and professional portfolios. These are intended for source verification and documentation. The README correctly warns that these URLs are unverified, demonstrating a security-conscious approach to external references.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 11:39 AM