The Agent Skills Directory

[SAFE]: The skill utilizes official PatSnap MCP tools (patsnap_search) and API endpoints (connect.patsnap.com) for patent data retrieval and AI chat functionality. All network operations are conducted within the vendor's verified ecosystem.
[SAFE]: The generator script (generate.py) safely processes patent records and injects them into a provided HTML template. It uses standard environment variables and bindings for output management within the agent environment.
[SAFE]: The generated platform allows for optional user-provided API Keys, which are stored in the browser's localStorage. The documentation includes appropriate security warnings, advising the use of backend proxies for production environments.
[SAFE]: External dependencies, such as Chart.js, are sourced from the trusted jsDelivr CDN, which is a well-known service for providing official library assets.
[SAFE]: The skill features an indirect data ingestion surface where external patent data is rendered in the dashboard UI. Given the trusted source (official PatSnap database) and the skill's primary purpose, this is considered a standard and acceptable behavior.
Ingestion points: Patent data is retrieved in generate.py and rendered via JavaScript in references/platform_template.html.
Boundary markers: Delimiters are not used for patent data interpolation.
Capability inventory: The Python script has file-write permissions for the generated HTML; the resulting browser application performs network fetch calls.
Sanitization: The generated HTML template renders patent metadata (titles/applicants) directly into the DOM without additional sanitization filters.

inner-mongolia-energy-ip-platform