The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where the agent ingests documentation, prop schemas, and test violation data from Storybook to guide automated code generation and 'self-healing' loops (documented in SKILL.md and rules/mcp-test-runner.md).
Ingestion points: Untrusted data enters the agent context through tools such as list-all-documentation, get-documentation, and run-story-tests which query the project's Storybook server.
Boundary markers: The instructions lack explicit boundary markers or directions to treat documentation content as data rather than instructions, potentially allowing embedded prompts to influence agent behavior.
Capability inventory: The skill uses Write and Edit tools (referenced in rules/mcp-test-runner.md) to modify source code, creating a path for ingested data to influence filesystem modifications.
Sanitization: No validation or sanitization mechanisms are specified for the component documentation or error messages processed during the self-healing cycle.
[COMMAND_EXECUTION]: Provides instructions to execute shell commands for upgrading Storybook and registering MCP servers using npx (e.g., npx storybook add @storybook/addon-mcp in SKILL.md).
[EXTERNAL_DOWNLOADS]: Fetches official Storybook addons and configurations from the NPM registry and the Chromatic cloud platform (https://chromatic.storybook.cloud) during initial setup and remote registration.

storybook-mcp-integration