The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a 'Self-heal' pattern in rules/mcp-test-runner.md that creates a vulnerability surface. The agent is instructed to ingest output from the run-story-tests tool (specifically failure details and accessibility violations) and use that data to automatically modify source code using the Edit or Write tools.
Ingestion points: Tool output from run-story-tests as described in rules/mcp-test-runner.md.
Boundary markers: No explicit delimiters or warnings to ignore embedded instructions within the test results are present.
Capability inventory: The skill utilizes Write and Edit tools to perform automated file modifications based on the ingested data.
Sanitization: No sanitization or validation logic is specified for the ingested test failure data before it influences code generation.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install official packages from the Storybook ecosystem.
Fetches the @storybook/addon-mcp and @storybook/mcp packages via npx and npm as documented in SKILL.md.
References the https://chromatic.storybook.cloud registry, which is an established service for Storybook hosting.
[COMMAND_EXECUTION]: The documentation includes shell commands required to set up the Storybook environment and register the MCP server.
Instructions include execution of npx storybook@latest upgrade, npx storybook add, and npx mcp-add for project configuration.

storybook-mcp-integration