The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads and installs the official Storybook CLI and various development addons through npx and npm. These resources are provided by the well-known Storybook organization and are considered safe for standard development workflows.
[REMOTE_CODE_EXECUTION]: The skill executes remote code by running npx storybook ai setup and npx storybook@latest init. These commands fetch and run the latest official Storybook setup logic. This behavior is considered safe as it targets a well-known and established service.
[COMMAND_EXECUTION]: The skill uses shell commands to analyze the project environment and install software. It also automates the registration of a Storybook MCP server by executing claude mcp add, which provides the agent with specific tools to interact with the Storybook environment.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection by processing project files to determine configuration settings.
Ingestion points: The scripts/discover-runtime.py script reads local project files including package.json, index.html, and application source code to identify React providers, CSS imports, and portal targets.
Boundary markers: No explicit delimiters are used to isolate data extracted from project files when it is interpolated into the generated .storybook/preview.tsx configuration.
Capability inventory: The skill possesses the Write, Edit, and Bash capabilities, which are used to generate and modify the project's Storybook configuration and execute setup scripts.
Sanitization: The discovery script employs regular expressions to extract specific patterns (such as component names), providing basic structural validation, but no further sanitization is performed on the data before it is used in code generation.

sb-setup