openui
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides detailed and secure implementation guidance for the OpenUI framework, including project scaffolding and Cloud integration. No malicious code or patterns were identified.
- [PROMPT_INJECTION]: The skill addresses potential indirect prompt injection by instructing the agent to treat remote content from official documentation and source code as reference data only. Ingestion points: https://github.com/thesysdev/openui and https://www.openui.com. Boundary markers: Explicit instructions are provided in SKILL.md. Capability inventory: The skill utilizes shell commands (npx, npm, pnpm) and file system access for configuration. Sanitization: Relies on model instruction to ignore embedded commands.
- [DATA_EXFILTRATION]: The runbooks provide clear security boundaries, mandating that THESYS_API_KEY remains on the server and that authenticated user IDs are derived from secure server-side session state.
- [EXTERNAL_DOWNLOADS]: All external packages and remote URL references belong to the official vendor (thesysdev) and are consistent with the skill's purpose.
- [COMMAND_EXECUTION]: The skill uses standard CLI tools (npx, npm, pnpm) for scaffolding and project management, following normal development workflows.
Audit Metadata