fsi-fa-check-deck
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to process document files and run its internal audit script. Evidence: Commands found in SKILL.md include
python -m markitdownfor document conversion andpython scripts/extract_numbers.pyfor data analysis. - [PROMPT_INJECTION]: The skill processes untrusted data from presentation files, creating a surface for indirect prompt injection.
- Ingestion points: Data from
presentation.pptxis converted to markdown and processed inSKILL.md. - Boundary markers: None present; the agent analyzes the extracted content directly without specific markers.
- Capability inventory: The skill utilizes subprocess execution and file system access across its scripts.
- Sanitization: The provided Python script filters numerical data, which reduces the surface for instructions embedded in numbers, but the agent also reads raw narrative text for language polishing.
Audit Metadata