pitch-deck-reviewer
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from pitch decks and has access to potentially exploitable tools like Bash and WebSearch.\n
- Ingestion points: Reads slide content from multiple formats including direct file reads (PDF, PPTX) and web fetching via links (SKILL.md).\n
- Boundary markers: No delimiters or safety instructions are present to distinguish between the skill's logic and the untrusted data it processes.\n
- Capability inventory: The skill uses Bash, Write, Edit, WebSearch, and WebFetch tools (SKILL.md).\n
- Sanitization: There is no evidence of input validation or content sanitization for the data extracted from the pitch decks.
Audit Metadata