brand-voice
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-supplied copy samples and data from platform connectors (e.g., call recordings, support logs) to derive brand attributes.
- Ingestion points: External copy samples and optional data connectors provided by the user.
- Boundary markers: Explicit instructions to present the user with an exact 'before/after' comparison of changes before commitment.
- Capability inventory: Capability to update the /foundation/brain.md file and log sessions to /context/skill-sessions.md.
- Sanitization: Risk is mitigated through mandatory user verification and confirmation of all generated content before it is saved.
- [DATA_EXFILTRATION]: The skill reads from and writes to specific local project files (/foundation/brain.md, /context/meta-patterns.md, /context/skill-sessions.md) to maintain brand context. All file operations are localized to the user's project environment and are standard for the skill's stated purpose. There is no evidence of network-based exfiltration or access to sensitive system credentials.
Audit Metadata