The Agent Skills Directory

[SAFE]: The skill uses environment variables (KIPRIS_API_KEY) for API authentication, which is the recommended method for preventing hardcoded credential exposure.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources during patent searches.
Ingestion points: Patent descriptions, claims, and metadata retrieved from the KIPRIS API through tools like patent_detail_search and patent_free_search (documented in SKILL.md).
Boundary markers: None explicitly provided to separate retrieved patent text from agent instructions.
Capability inventory: The skill writes retrieved data to local Excel files via output_path parameters.
Sanitization: No specific sanitization or filtering of API responses is mentioned before file writing or analysis.
[SAFE]: The workflow involves standard data processing tasks, such as batch exporting results and deduplicating records based on application numbers, which align with the skill's stated purpose.

patent-search-collect