The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of Playwright and the subsequent download of Chromium browser binaries (playwright install chromium) to support its image rendering features.
[CREDENTIALS_UNSAFE]: The skill processes sensitive Xiaohongshu session cookies (XHS_COOKIE). The publish_xhs.py script contains an 'API mode' that transmits these credentials over the network to a configurable XHS_API_URL. While the default is localhost, this capability creates a potential exposure vector for session tokens if the endpoint is redirected or misconfigured.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the following evidence chain:
Ingestion points: User-provided materials are ingested and converted into Markdown documents to be used as templates for image generation.
Boundary markers: Absent; the rendering pipeline does not use delimiters or instructions to ignore embedded code in the Markdown content.
Capability inventory: The skill uses Playwright to open and render HTML in a browser environment and can perform network POST requests to external APIs.
Sanitization: Absent; the conversion from Markdown to HTML in scripts/render_xhs.py and scripts/render_xhs.js does not include sanitization, allowing potential XSS payloads to execute within the local headless browser.
[COMMAND_EXECUTION]: The SKILL.md instructs the AI agent to execute local Python and Node.js scripts using shell commands to perform the core functions of the skill.

xhs-note-creator