brand-voice
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill creates a 'voice profile' by analyzing untrusted external data provided by users, such as writing samples, emails, or content fetched from URLs.
- Ingestion points: Samples are gathered via user input or external URLs in the 'Initial assessment' phase.
- Boundary markers: No specific boundary markers or 'ignore embedded instructions' warnings are implemented when processing the samples.
- Capability inventory: The resulting analysis is written to project files (
.agents/blog-context.md) which directly influence the behavior of downstream skills like 'article' and 'humanize'. - Sanitization: There is no mention of sanitizing or validating the content of the samples before they are integrated into the persistent profile.
Audit Metadata