The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data (patent titles, abstracts, and summaries) from external Patsnap API endpoints and incorporates them into generated Markdown and HTML reports. This surface is mitigated by explicit sanitization logic.
Ingestion points: scripts/fetch_competitor_report.py (via search, get_patent_detail, get_ai_summary) and scripts/fetch_literature.py.
Boundary markers: The skill instructions mandate separating retrieved facts from AI interpretation and marking unverified data explicitly.
Capability inventory: The skill possesses file-write capabilities (reports/ directory) and network access via the requests library.
Sanitization: The scripts/render_html.py script uses html.escape for all text fields and implements a whitelist for URL schemes and image paths via _safe_url.
[COMMAND_EXECUTION]: The skill provides a shell script scripts/run.sh as an entry point to coordinate the Python-based data retrieval and report rendering pipeline. The script includes logic to sanitize the report title before using it in a filename.
[EXTERNAL_DOWNLOADS]: The skill fetches patent metadata and binary images from Patsnap's official API domains (open.patsnap.com, analytics.patsnap.com). These downloads are integral to the skill's primary purpose and target the vendor's own infrastructure.
[DYNAMIC_EXECUTION]: The Python scripts utilize sys.path.insert to manage local module imports within the skill directory, which is a standard practice for multi-file script packages.

run-advanced-patent-query-ip