The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands for HTML validation and testing.
Evidence: The skill instructs the agent to execute python3 -m html.parser and rg (ripgrep) commands on generated reports to verify link integrity and HTML structure.
Evidence: Mentions using Playwright to test the interactive hover functionality in the generated reports.
[DYNAMIC_EXECUTION]: The skill generates HTML reports containing embedded JavaScript for user interface enhancements.
Evidence: references/report-spec.md provides a JavaScript IIFE (Immediately Invoked Function Expression) to be embedded in reports for fixed-position image hover previews.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by ingesting untrusted data from external patent and news sources into HTML templates.
Ingestion points: Data enters the context via ls_* and patsnap_* MCP tools (patent data, news, clinical trials) in SKILL.md.
Boundary markers: Absent; data is interpolated directly into V6-style HTML structures.
Capability inventory: The skill has file-write access to the reports/ directory and can execute shell commands (rg, python3) for validation.
Sanitization: Not explicitly defined for the content of the patent data being processed.
[EXTERNAL_DOWNLOADS]: The skill references external dependencies and services for data retrieval and testing.
Evidence: Usage of Patsnap/智慧芽 MCP tools for fetching real-time data from zhihuiya.com domains.
Evidence: Mentions Playwright as a testing dependency for verifying HTML report functionality.

zhihuiya-patent-html-report