headless-component
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- Standard Command Execution: The skill directs the agent to execute development commands via the Bash tool, including yarn nx for builds and tests, yarn beachball for change management, and gh stack for handling stacked pull requests. These operations are restricted to the local development environment and the project's official GitHub repository.
- File System Operations: The instructions involve reading and modifying source code, manifests, and documentation within the repository's package structure. This is the primary function of the skill and is handled through established tools like Read, Write, and Edit.
- Indirect Prompt Injection Surface: The skill involves the agent reading existing source code and RFC documents to inform its work. While this is necessary for context, it presents a theoretical surface where instructions embedded in repository files could influence agent behavior. The skill includes specific validation steps—such as verifying PascalCase for component names—which act as a basic constraint on input processing.
- Network Interaction: The use of the gh (GitHub) CLI facilitates interaction with remote repositories for code submission. The instructions provide explicit guidance on selecting the correct remote (e.g., microsoft/fluentui vs. a personal fork) to ensure code is pushed to the intended destination.
Audit Metadata