The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands to verify the existence of components and documentation directories within the project repository (e.g., searching apps/docs/docs/components/ and packages/web/src/).- [COMMAND_EXECUTION]: Instructs the agent to perform automated directory and file creation for component documentation using standard bash commands.- [COMMAND_EXECUTION]: The skill modifies project source code and configuration files, such as ReactLiveScope.ts, sidebars.ts, and docgen.config.js, to register and integrate new component documentation into the site.- [SAFE]: The skill performs research by querying well-known and reputable third-party documentation sites (Material UI, Radix UI, Mantine, Ant Design, and Base UI) to ensure the generated documentation follows industry-standard patterns.- [SAFE]: Indirect Prompt Injection Surface: The skill ingests data from external documentation websites during the research phase. While these are reputable targets, the agent uses this information to generate project files and modify configurations. There are no explicit boundary markers or sanitization steps for the ingested web content; however, the risk is assessed as negligible due to the specific, technical nature of the task and the high reputation of the target domains.- [SAFE]: The skill accesses local project metadata (e.g., package.json, IconName.ts) to extract valid configuration values and peer dependencies, which is appropriate for its stated purpose of documentation generation.

components.write-docs