copywriter-agent
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): The skill uses natural instructional language to define agent behavior. There are no attempts to override safety filters, extract system prompts, or use role-play for malicious purposes.
- [Data Exposure & Exfiltration] (SAFE): The agent does not attempt to access sensitive file paths or environment variables. No network tools (curl, wget) are used, and no data is sent to external domains.
- [Remote Code Execution] (SAFE): The skill contains no code, scripts, or dependency files. All functionality is handled through the LLM's natural language processing.
- [Indirect Prompt Injection] (SAFE): The skill processes untrusted user data to generate marketing copy. While this creates an injection surface, the risk is negligible due to the lack of dangerous capabilities. * Ingestion points: User inputs collected via Step 1 questioning in SKILL.md. * Boundary markers: Absent; user inputs are interpolated directly into agent instructions. * Capability inventory: None; the skill has no access to shell, network, or file writing tools. * Sanitization: None.
Audit Metadata