The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data via command-line arguments and stores it in persistent files, which could be exploited to influence the agent's behavior during future interactions with those files.
Ingestion points: Data provided via the --asset and --subtitle arguments in SKILL.md is interpolated into JSON and HTML files.
Boundary markers: There are no specified delimiters or "ignore" instructions for the content written to design-assets.json or assets.html to prevent the agent from treating stored data as instructions.
Capability inventory: The skill utilizes the Write tool to modify files and Bash to execute local scripts, providing a pathway for malicious data to persist in the workspace.
Sanitization: No explicit sanitization, validation, or escaping of the user-provided strings is described before they are stored or rendered.
[DYNAMIC_EXECUTION]: The skill invokes a local Node.js script to process the design data and regenerate the user-facing asset grid.
Execution method: The command node scripts/make-assets-index.mjs is executed via the Bash tool using a static local path.

register-asset