create-apple-notes-mockup
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted JSON data which is interpolated into an HTML template. Ingestion points: Input JSON file passed via the --note flag in render.js. Boundary markers: No explicit markers provided; the tool functions as a deterministic renderer. Capability inventory: File system read/write via fs module and browser automation via Playwright. Sanitization: Implements an escapeHtml function in generate.js that sanitizes text fields (title, paragraphs, checklist items, keyboard suggestions) before they are inserted into the HTML document.
- [EXTERNAL_DOWNLOADS]: The skill requires downloading the Chromium browser using the official Playwright installation command. This is a standard and expected procedure from a well-known technology provider.
- [COMMAND_EXECUTION]: The skill executes local Node.js scripts and browser automation tasks that are essential to its documented functionality for rendering screenshots.
- [DATA_EXFILTRATION]: The render.js script resolves image paths relative to the input file. This creates a potential surface where a malicious input could cause the tool to read local sensitive files and embed them in the output as Base64 data. However, the skill provides no automated network exfiltration mechanism.
Audit Metadata