The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes instructions to install the hf-gradio extension from the gradio-app/hf-gradio repository using the hf CLI. This targets a repository owned by the skill's author.
[COMMAND_EXECUTION]: The skill utilizes the gradio, hf-gradio, and hf gradio CLI commands to perform administrative and predictive tasks against remote Gradio applications.
[DATA_EXFILTRATION]: The predict command functionality allows for local files to be read and uploaded to remote servers (Hugging Face Spaces or public URLs) when the gradio.FileData metadata is included in the JSON payload. This is a documented feature of the Gradio API intended for processing local file inputs.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by ingesting and processing JSON responses from external, potentially untrusted Gradio applications.
Ingestion points: Data returned by the info and predict commands from remote endpoints (SKILL.md).
Boundary markers: No specific delimiters or safety instructions are provided to the agent regarding the handling of remote API responses.
Capability inventory: The skill possesses the capability to perform network operations via the gradio CLI and read local files for transmission to remote endpoints (SKILL.md).
Sanitization: No explicit sanitization or validation of the remote API response is defined within the skill's usage instructions.

hf-gradio