pitch-deck
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The workflow specifies the use of shell commands including
soffice(LibreOffice) andpdftoppmto convert PowerPoint presentations into PDF and then into images for visual validation purposes. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and process data from untrusted external source materials.
- Ingestion points: Data is extracted from Excel, CSV, PDF reports, Word documents, databases, and web sources as part of Phase 1 and Phase 2 of the workflow.
- Capability inventory: The agent performs file system operations (read/write PPTX files) and executes shell commands for file conversion.
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are utilized when interpolating source data into the template context.
- Sanitization: The instructions do not define sanitization, escaping, or filtering procedures for the content extracted from external source materials.
Audit Metadata