The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes Python scripts (ruiguan_trademark_graphic_detection.py and upload_image.py) to handle API communication and image processing. These scripts are executed as subprocesses with parameters provided by the agent.
[EXTERNAL_DOWNLOADS]: The skill connects to tool-gateway.linkfox.com for trademark detection queries and image storage. It also suggests downloading an onboarding skill from agent-files.linkfox.com if authentication fails. These domains are owned by the skill's author.
[DATA_EXFILTRATION]: The upload_image.py script is designed to read local files and upload them to LinkFox cloud storage. While intended for product images, this mechanism facilitates the movement of local file content to a remote server as part of its primary functional requirement.
[PROMPT_INJECTION]: The skill processes and displays external trademark data from a remote database, creating a potential indirect prompt injection surface.
Ingestion points: Trademark database results and radar assessment strings from the Ruiguan API.
Boundary markers: Results are rendered into tables or summaries without explicit delimiters between the external data and agent instructions.
Capability inventory: The associated scripts have the ability to read local files and write results to the workspace.
Sanitization: No specific string sanitization or escaping logic is implemented for the data retrieved from the API.
[SAFE]: The skill manages authentication using environment variables (LINKFOX_AGENT_API_KEY), which is a secure practice for handling API credentials.

linkfox-ruiguan-trademark-graphic-detection