design-animate
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input through the $ARGUMENTS variable and by scanning project files without explicit boundary markers or sanitization. Combined with its 'zero questions' autonomous mandate and code-writing capabilities, this presents a surface for malicious instructions to influence the agent's output.
- Ingestion points: Project UI files, configuration files (package.json, pubspec.yaml), and the $ARGUMENTS input field.
- Boundary markers: None present; data and instructions are treated uniformly.
- Capability inventory: File system read/write access across the project and execution of shell commands for build verification.
- Sanitization: No sanitization or validation of input data is performed.
- [COMMAND_EXECUTION]: The instructions require the agent to autonomously execute shell commands, specifically project build commands, and perform file system modifications (CSS, JavaScript, and Dart files) without user confirmation or questioning.
Audit Metadata