The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or injection attempts were detected in the skill's instructions, metadata, or scripts.
[DATA_EXFILTRATION]: Network activity is confined to the vendor's domains (linkfox.com) for patent data retrieval and feedback submission. These operations are transparently documented and consistent with the skill's intended functionality.
[CREDENTIALS_UNSAFE]: The skill implements safe secret management by retrieving the necessary API key from the LINKFOXAGENT_API_KEY environment variable rather than using hardcoded values.
[COMMAND_EXECUTION]: The execution logic in scripts/zhihuiya_patent_family.py is limited to making a single POST request and parsing the JSON response. It does not utilize subprocesses or dynamic code execution functions like eval() or exec().
[EXTERNAL_DOWNLOADS]: The skill does not download or execute any external scripts or packages during runtime. It relies entirely on Python's standard library.

linkfox-zhihuiya-patent-family