hypogenic
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Prompt Injection (LOW): The configuration defines an attack surface for indirect prompt injection by interpolating untrusted external data into LLM prompts.
- Ingestion points: Reads from data/train.json, data/val.json, data/test.json, and PDF files in the papers/ directory.
- Boundary markers: Absent. The templates (e.g., {data_samples}, {sample_text}) do not use delimiters or instructions to ignore instructions found within the data.
- Capability inventory: LLM API interaction and file writing to the output/ directory.
- Sanitization: No input validation or sanitization logic is defined in the template.
- Credentials (SAFE): The template uses api_key_env: "OPENAI_API_KEY", which is a secure practice to avoid hardcoding API keys.
Audit Metadata