The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied dashboard requests to generate YAML, TSX, and SQL files. This creates an attack surface where malicious input could influence the structure or content of the generated dashboard components or queries.
Ingestion points: [dashboard request] argument hint; user-provided data for filters and widgets.
Boundary markers: Not explicitly provided in instructions for file generation.
Capability inventory: The skill utilizes tool commands (dac validate, dac check, dac serve) and performs file operations to create dashboard projects.
Sanitization: The documentation notes that runtime URL parameters are validated against types, but doesn't specify sanitization for the agent's generation process.
[DYNAMIC_EXECUTION]: The skill generates TSX (TypeScript React) files and YAML definitions that are executed or rendered by the dac CLI tool. This dynamic generation and execution of code is central to the skill's purpose but represents a standard risk for code-generation utilities.
[COMMAND_EXECUTION]: The instructions involve the use of the dac CLI tool for project initialization, validation, and serving. These commands are necessary for the development workflow and operate on the local project directory.

create-dashboard