creating-concepts
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions designed to influence the agent's behavior during document creation.
- Ingestion points: Raw workshop material including notes, transcripts, and whiteboard photo transcriptions are processed to create document content (reference/concept-structure.md).
- Boundary markers: There are no explicit instructions to use delimiters or ignore embedded commands when processing these sources.
- Capability inventory: The skill has access to tools that can modify platform data (
create_offer,update_offer) and perform state changes (mark_sent) based on the processed content (SKILL.md). - Sanitization: Although the platform is said to sanitize
printHtml, there is no evidence of sanitizing the raw text inputs used to build the main content blocks.
Audit Metadata