structured-output
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface identified.
- Ingestion points: The input_text parameter in scripts/generate_object.py and the input field in SKILL.md templates.
- Boundary markers: The prompt uses simple text headers like 'Input:' but lacks robust delimiters or specific instructions for the model to ignore instructions within the input.
- Capability inventory: The skill is focused on text processing and JSON generation; the included scripts do not have access to dangerous system capabilities or network tools.
- Sanitization: No sanitization or escaping is applied to the untrusted input before it is placed in the prompt.
Audit Metadata