gts-design-system-integration
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (HIGH): The skill creates a high-risk attack surface by ingesting untrusted application code and executing shell commands based on its analysis.
- Ingestion points: The agent reads and identifies the user's existing app structure and styling stack (SKILL.md).
- Boundary markers: No delimiters or protective instructions are used to separate untrusted code from agent instructions.
- Capability inventory: The skill directs the agent to execute shell commands including npm install, npm run build, and npm run lint (SKILL.md, references/guidelines.md).
- Sanitization: No validation or sanitization of the user-provided code is performed before command execution.
- [Unverifiable Dependencies & Remote Code Execution] (MEDIUM): The skill mandates the installation of gts-central-library@0.1.0-beta.2, an external package from an unspecified and non-trusted source.
- [Command Execution] (LOW): Explicit instructions to run developer tools (npm/bun build and lint) are included for validation, which can be exploited if the project's scripts are compromised.
Recommendations
- AI detected serious security threats
Audit Metadata