visual-style-ppt
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data including Markdown, TXT, PDF, Word, and PPT documents, as well as reference images, to generate presentation content. It lacks explicit boundary markers (e.g., delimiters like """ or XML tags with warnings) to isolate this untrusted content from the agent's internal instructions. This creates a vulnerability where malicious instructions embedded in a processed document could potentially influence the agent's behavior during the slide generation or packaging process.
- Ingestion points: The skill reads user-provided documents (Markdown, TXT, PDF, Word, PPT) and visual reference images to extract outlines and visual styles (SKILL.md, workflow.md).
- Boundary markers: There are no specific instructions or delimiters mentioned to separate processed data from agent instructions.
- Capability inventory: The skill has capabilities to read local files, write files (images, PPTX, and ZIP archives), and perform network operations via the Image 2 model API.
- Sanitization: The instructions do not define any sanitization, filtering, or escaping mechanisms for the text extracted from external documents before it is used to formulate Image 2 prompts or PPT content.
Audit Metadata