forecasting-reverso

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly downloads and/or accepts public HuggingFace checkpoint files (see SKILL.md "Obtaining Weights" Path A URL and the fallback upload path; scripts/reverso.py & scripts/load_checkpoint.py load and unpickle those .pth/.npz weights), which are untrusted third‑party/user‑provided content that the agent loads and uses to produce forecasts that can materially influence downstream decisions.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The setup commands include a system-level pip install with "--system --break-system-packages", which instructs modifying system packages/installation state (potentially breaking system-managed packages), so the skill pushes changes to the machine state.