alterlab-pptx-posters
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates the creation of posters by calling local system tools such as google-chrome in headless mode for PDF generation and libreoffice for PPTX conversion. It also provides a Python script snippet using the python-pptx library that the agent is expected to write and execute locally to produce PowerPoint files. Additionally, it invokes internal Python scripts to generate visual elements based on user input.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. User-provided data, such as research topics and section content, is interpolated directly into shell command strings used to call image generation tools. If the user provides crafted input, it could potentially manipulate the command execution or the prompts sent to subordinate AI tools. 1. Ingestion points: User-supplied research topics and poster text content processed in SKILL.md. 2. Boundary markers: The skill instructions recommend wrapping prompts in double quotes, but there are no specific delimiters or instructions to ignore nested commands within the user data. 3. Capability inventory: The skill has access to shell execution via Bash and the ability to write files to the local system. 4. Sanitization: No sanitization or validation of user-provided strings is performed before they are used to build shell commands.
Audit Metadata