The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to generate and execute shell commands using curl, jq, and the hf CLI tool. These commands target the official huggingface.co API and repository infrastructure to perform tasks like downloading model cards and querying service metadata.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from the Hugging Face Hub, including model/dataset cards, repository metadata, and API responses. Because these sources contain user-provided text, they represent an attack surface where malicious instructions could be embedded to influence the agent's behavior during script creation or data processing.
Ingestion points: Hugging Face API endpoints (e.g., /api/models, /api/datasets) and repository content (Model/Dataset cards) accessed via the hf CLI or curl (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" blocks for handling the ingested external data.
Capability inventory: Generation and execution of shell scripts (bash), Python scripts, and TSX executables; network access to well-known domains via curl; and local file management via the hf CLI.
Sanitization: No explicit sanitization, filtering, or escaping mechanisms are mandated for external content before it is processed or used in script generation.

hugging-face-tool-builder