torchforge-rl-training
Warn
Audited by Snyk on Jun 29, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (medium risk: 0.65). Runtime GRPO training ingests dataset examples from a user-specified external dataset path (e.g.,
openai/gsm8kvia HuggingFace streaming), so the LLM context can include outsider-authored free-form text from that dataset as prompts/responses.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata