brand-guidelines
Warn
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: MEDIUMNO_CODESAFE
Full Analysis
- [METADATA_POISONING]: The YAML frontmatter in SKILL.md lists the author as 'Anthropic', which contradicts the identified author 'zocomputer'. This represents a misleading claim of provenance and impersonation of a well-known entity.
- [SAFE]: The skill contains no executable code, scripts, or binary assets. It is strictly informational documentation.
- [SAFE]: No network operations, file system access, or credential harvesting patterns were identified.
- [SAFE]: The instructions do not contain prompt injection patterns or attempts to override the agent's core safety behavior.
Audit Metadata