curriculum-develop-items
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [Category 1: Prompt Injection] (SAFE): No instructions found that attempt to override system prompts, bypass safety filters, or extract system instructions. The instructions are focused on educational content creation.
- [Category 2: Data Exposure & Exfiltration] (SAFE): No sensitive file paths (e.g., SSH keys, credentials), hardcoded secrets, or network calls (curl, wget) were detected. The file references are local curriculum artifacts.
- [Category 4: Unverifiable Dependencies & Remote Code Execution] (SAFE): No external package installations or remote script executions. The 'CLI Interface' section describes internal agent commands rather than shell commands to be executed on a host system.
- [Category 8: Indirect Prompt Injection] (LOW):
- Ingestion points: Processes external 'Assessment Blueprints' and 'Learning Objectives'.
- Boundary markers: None explicitly defined in the provided file.
- Capability inventory: No dangerous capabilities like subprocess execution, file writes, or network operations.
- Sanitization: None mentioned, but the risk is inherently low as the output is static assessment text.
Audit Metadata