html-presentation-deck
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data (outlines, documents, PPTX) to generate slide content, creating a potential surface for indirect prompt injection.
- Ingestion points:
src/slides.tsx,content/slides.json, and various user-provided source materials. - Boundary markers: Absent; there are no explicit delimiters or instructions to disregard embedded commands in the source material.
- Capability inventory: The agent has the ability to write files and execute shell commands (
npm run build,npm run preview). - Sanitization: No sanitization or validation logic for external content is described.
- [COMMAND_EXECUTION]: Instructs the creation and execution of local launcher scripts (
打开预览.cmd,preview.ps1) to facilitate local project previews and utilizes standard CLI tools (npm,vite,gh) for build and deployment processes. - [EXTERNAL_DOWNLOADS]: Fetches design guidelines and baseline visual references from the vendor's official GitHub Pages repository (
water-nn.github.io).
Audit Metadata