marketing-copy
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): The instructions are focused on copywriting style and principles. No patterns were found that attempt to override system instructions or bypass safety filters.
- Data Exposure & Exfiltration (SAFE): No hardcoded secrets, sensitive file paths, or network-based exfiltration vectors were detected in the prompt or metadata.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill consists entirely of markdown and JSON metadata. It does not include scripts (Python, JS, etc.) or commands to download/execute external code.
- Indirect Prompt Injection (LOW):
- Ingestion points: The skill accepts user-provided product names, target audiences, and core transformations in
SKILL.md. - Boundary markers: Absent; the prompt uses simple placeholders like
[PRODUCT/SERVICE]. - Capability inventory: None; the skill has no ability to write files, execute commands, or access the network.
- Sanitization: Absent; however, given the lack of system capabilities, the risk of exploiting these inputs is negligible.
Audit Metadata