infographic-structure-creator
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- Indirect Prompt Injection (HIGH): The skill is designed to ingest data from external files and the local file system to guide code generation, creating an exploitable capability tier.
- Ingestion points: The workflow requires reading 'references/structure-prompt.md' and scanning 'src/designs/structures'.
- Boundary markers: Absent. There are no instructions to delimit data from instructions or to ignore embedded commands in the source files.
- Capability inventory: The skill is explicitly tasked with producing full TypeScript/TSX files and modifying the 'src/designs/structures' directory, which constitutes a significant write capability.
- Sanitization: Absent. No validation logic is present to ensure the generated code does not contain malicious logic or unauthorized side effects.
- Dynamic Execution (MEDIUM): The skill generates executable script files (.tsx) based on locally-generated content and external references, which is a form of script generation that will be compiled and executed by the host environment.
Recommendations
- AI detected serious security threats
Audit Metadata