brand-guidelines
Pass
Audited by Gen Agent Trust Hub on Apr 11, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It instructs the agent to style 'any sort of artifact', which serves as an ingestion point for untrusted data that could contain hidden instructions.\n
- Ingestion points: Documented in
SKILL.mdas 'any sort of artifact that may benefit from having Anthropic's look-and-feel'.\n - Boundary markers: Absent; there are no instructions provided to use delimiters or to ignore embedded content within the ingested artifacts.\n
- Capability inventory: No scripts are included in the skill, but
SKILL.mdreferences the use of document manipulation capabilities (specificallypython-pptx) to apply styles.\n - Sanitization: No validation or sanitization of the input artifact content is specified.\n- [NO_CODE]: The skill contains only documentation, metadata, and licensing information (
SKILL.md,LICENSE.txt) and does not include any executable scripts or binaries.
Audit Metadata