z-wanghong-handwritten-ppt
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides scripts (
scripts/render.sh,scripts/prepare_render_html.py) that use the local Python interpreter and Google Chrome binary to automate the conversion of HTML templates into PNG images. These tools operate on local files created by the agent during the task. - [PROMPT_INJECTION]: The skill is designed to process user-supplied text to generate HTML presentation slides, which constitutes an indirect prompt injection surface.
- Ingestion Points: User-provided articles or technical summaries used to populate the PPT templates.
- Boundary Markers: Not implemented; user content is interpolated directly into HTML sections.
- Capability Inventory: The skill can execute local scripts and trigger a headless browser session to render the generated HTML via
scripts/render.sh. - Sanitization: Content is rendered in a browser environment without explicit sanitization, which could lead to code execution if malicious HTML or script tags are provided in the source text.
- [SAFE]: The skill utilizes local system resources such as the 'HanziPen SC' font and standard browser binaries to perform its primary function. No network exfiltration, unauthorized file access, or suspicious remote dependencies were observed.
Audit Metadata