components-write-docs
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, including component source code, Storybook files, and web search results from third-party documentation sites. This content is used to generate documentation and modify project files.
- Ingestion points: The agent reads component source files (
packages/web/src/...), package configuration (package.json), and external content via web search (e.g., Material UI, Radix UI). - Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the agent when processing these external inputs.
- Capability inventory: The skill can execute shell commands (
bash), write new documentation files, and modify existing TypeScript and JavaScript configuration files (ReactLiveScope.ts,sidebars.ts,docgen.config.js). - Sanitization: No explicit sanitization or filtering of the ingested external content is described.
- [DYNAMIC_EXECUTION]: The skill generates and modifies build-critical configuration files such as
docgen.config.js,sidebars.ts, andReactLiveScope.ts. While these actions are central to its purpose of automating documentation setup, programmatically assembling executable or configuration code based on external inputs represents a potential attack surface. - [COMMAND_EXECUTION]: The skill uses shell commands to verify the existence of files and directories within the repository structure (e.g., checking for existing documentation or component source files).
- [EXTERNAL_DOWNLOADS]: The skill directs the agent to perform web searches for research purposes, specifically targeting well-known UI component libraries like Material UI, Radix UI, Mantine, Ant Design, and Base UI to inform documentation structure and best practices.
Audit Metadata