The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its core functionality.
Ingestion points: The skill ingests untrusted external data via web.search and patent.search results (documented in Module 1, 2, and 3) to generate analysis reports.
Boundary markers: The instructions do not specify explicit delimiters or boundary markers to isolate the retrieved external content from the agent's internal processing logic.
Capability inventory: The skill possesses the capability to write HTML reports to the @session/ directory and generate/download Word documents using the office tool suite.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from external web searches before it is interpolated into the report generation prompts.
[EXTERNAL_DOWNLOADS]: The skill references several external resources and specialized Model Context Protocol (MCP) tools, including patsnap-find-solution and novelty-search-lite. It also communicates with connect.zhihuiya.com (the official Patsnap API endpoint). These are verified vendor resources belonging to the author, Patsnap, and are used for legitimate business functionality.
[CREDENTIALS_UNSAFE]: The skill provides placeholders and instructions for users to input their own 'Patsnap Open Platform API Key'. These are standard configuration prompts and do not represent hardcoded or leaked credentials.

rd-ip-accelerator