writing-ad-copy

Pass

Audited by Gen Agent Trust Hub on Aug 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or dangerous execution capabilities were detected. The skill is limited to generating text-based ad copy based on user-supplied information and predefined platform guidelines.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (such as customer reviews and ad comments) which presents a potential surface for indirect prompt injection.
  • Ingestion points: Grounding material (reviews, comments, and customer language) is ingested in SKILL.md (Step 1) to build ad variants.
  • Boundary markers: The instructions do not provide explicit delimiters or warnings to ignore embedded instructions within the user-supplied grounding material.
  • Capability inventory: The skill has no capabilities (no allowed tools, subprocesses, or network access) to perform actions outside of returning text to the user.
  • Sanitization: No explicit sanitization or filtering of external input is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 22, 2026, 05:35 PM
Security Audit — agent-trust-hub — writing-ad-copy