The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute belt CLI commands, which is the intended method for interacting with the Inference.sh platform.
[REMOTE_CODE_EXECUTION]: Python and HTML code snippets are sent to remote executors (infsh/python-executor and infsh/html-to-image) via the belt CLI to generate visualization assets.
[EXTERNAL_DOWNLOADS]: The skill references the installation of the belt-sh/cli package and points to documentation hosted on the vendor's official GitHub repository.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of charts from data inputs, which represents a potential attack surface for indirect prompt injection if an agent is tasked with visualizing untrusted external content.
Ingestion points: Data values, labels, and titles provided in the --input JSON payload for belt commands in SKILL.md.
Boundary markers: No explicit delimiters or safety instructions are used to separate data from the generated code strings.
Capability inventory: Includes shell command execution and remote code execution capabilities.
Sanitization: The example scripts do not demonstrate sanitization or escaping of input data before its interpolation into executable Python or HTML code.

data-visualization