The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes the 'current topic' to generate HTML artifacts. This topic is derived from the agent's context or user input, which may be untrusted.
Ingestion points: The current topic or user-provided description in SKILL.md.
Boundary markers: There are no explicit instructions to use delimiters or ignore embedded instructions within the data being visualized.
Capability inventory: The skill performs file writing (implied by creating .html files) and uses the open command.
Sanitization: There are no instructions to sanitize or escape HTML/JavaScript content before generating and opening the artifact.
[COMMAND_EXECUTION]: The skill instructs the agent to execute the open command on generated HTML files (e.g., open show-me-{description}.html). While this is consistent with the skill's purpose, it executes a shell command on a filename and content derived from potentially untrusted input.

show-me