copy-pua
Pass
Audited by Gen Agent Trust Hub on Jun 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses a structured role-play approach ("Mean Editor") and specific corporate personas (e.g., Alibaba, ByteDance, Jobs) to steer the agent's tone during the editing process. This is intended functionality and does not attempt to bypass underlying model safety filters.
- [SAFE]: Data access is limited to a local configuration file (
.copy-pua/writing-dna.md) used to store user preferences and writing styles. The instructions include explicit warnings for users to avoid committing sensitive information to public repositories. - [SAFE]: The skill does not perform network exfiltration, hardcode credentials, or execute arbitrary commands. The "hooks" described in the documentation are optional manual configurations for the user's host environment.
- [SAFE]: The skill instructions are self-contained and do not rely on unverifiable external dependencies or remote script execution.
Audit Metadata