The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute local Python scripts (scripts/fetch_articles.py and scripts/generate_hot_html.py) to retrieve data and generate reports. These scripts use the subprocess module to perform their tasks.
[DATA_EXPOSURE]: Inside scripts/fetch_articles.py, the script explicitly disables SSL certificate verification by setting ssl_ctx.verify_mode = ssl.CERT_NONE. This makes the connection to the vendor's API vulnerable to Man-in-the-Middle (MITM) attacks, which could lead to the exposure of the REDFOX_API_KEY and the retrieved data.
[EXTERNAL_DOWNLOADS]: The HTML report generated by scripts/generate_hot_html.py includes a reference to the html2pdf.js library hosted on Cloudflare's well-known public CDN (cdnjs.cloudflare.com). This is used to provide PDF export functionality for the user.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests and displays content from an external API (redfox.hk).
Ingestion points: scripts/fetch_articles.py fetches article titles and metadata from an external JSON API.
Boundary markers: None. The agent is specifically instructed in SKILL.md to display the script's output "as-is" and "without modification."
Capability inventory: The agent has the capability to execute commands via subprocess and perform local file writes.
Sanitization: No sanitization or filtering of the content retrieved from the API is performed before it is rendered by the agent or included in the HTML report.

wechat-original-hot