Confirm the source
Compare the repository, publisher, and installation command before continuing.
✓
Third-party audit
PassedGen Agent Trust Hub
sepia-refactor · nanako0129/sepia
Provider summary
This skill acts as a secure wrapper for a local prose refactoring tool. It enforces strict relative pathing for its dependencies and includes explicit instructions for the AI to treat user input as untrusted data while denying all tool and network authorities, effectively mitigating common risks like prompt injection and data exfiltration.
The provider classified the analyzed snapshot as NONE risk. This result applies to the content available at audit time.
- Provider
- Gen Agent Trust Hub
- Normalized result
- pass
- Risk level
- NONE
- Audited
- Aug 30, 2026
How to use this signal
Read the content
Review SKILL.md, scripts, and allowed tools. OpenAI Suite never executes them during indexing.
Limit permissions
Use the narrowest possible scope and avoid broad credentials for third-party skills.