The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill's behavior is consistent with its stated purpose of patent data retrieval.
[EXTERNAL_DOWNLOADS]: The skill communicates with the vendor's API endpoint (NEXSCOPE_PROXY_BASE) to fetch patent claim data. This is the primary function of the skill and uses official channels.
[COMMAND_EXECUTION]: The script scripts/patent_claims_translation.py performs routine file system operations, including creating directories and writing JSON response files to the project's nexscope folder and the user's home directory for caching. These operations are transparent and follow standard development practices.
[INDIRECT_PROMPT_INJECTION]: The skill ingests patent claim text from the API, which serves as a data entry point from an external source.
Ingestion points: The claims field from the API response in scripts/patent_claims_translation.py.
Boundary markers: No explicit safety delimiters are added to the raw script output, although the SKILL.md defines structured display rules for the agent.
Capability inventory: The skill possesses file-write and network-read capabilities through the included Python script.
Sanitization: The retrieved patent text is passed to the agent without specific sanitization or filtering.

ecommerce-patent-claims-translation