copywriting
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides structured guidance for creating marketing copy using standard frameworks like AIDA and PAS. No security risks or malicious patterns were identified in the instructions or workflow.
- [PROMPT_INJECTION]: The instructions use emphatic language (e.g., 'MANDATORY', 'MUST ATTENTION') to enforce operational procedures such as task creation and evidence citing. These are process-oriented constraints and do not attempt to bypass safety filters, extract system prompts, or override agent guardrails.
- [DATA_EXFILTRATION]: No network operations, such as curl or wget, or access to sensitive file paths (e.g., .ssh, .env) were detected. The instruction to search the codebase is consistent with the agent's intended use-case for project-specific copywriting.
- [REMOTE_CODE_EXECUTION]: There are no commands for downloading external scripts, installing unverified packages, or executing dynamic code. The skill is entirely instructional and does not involve executable scripts.
Audit Metadata