slide-making-skill
Pass
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for incorporating external resources such as images and background textures via URLs into the generated slides using the PptxGenJS API.
- [COMMAND_EXECUTION]: The skill contains a QA section that instructs the agent to run command-line tools, specifically
python -m markitdown, to verify the output of generated PowerPoint files and suggests environment setup vianpm install. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes user-provided content to generate executable JavaScript files and performs verification via shell commands.
- Ingestion points: User input content intended for slide titles, body text, and chart data within the generated JavaScript files.
- Boundary markers: No specific delimiters or "ignore instructions" warnings are provided to separate user content from the generated code logic.
- Capability inventory: The agent generates code that performs file system writes (
pres.writeFile) and executes shell commands (python -m markitdown) during the verification loop. - Sanitization: The instructions do not specify any sanitization, escaping, or validation steps for user-provided data before it is interpolated into the executable script.
Audit Metadata