A safer way to give Hermes live data

Hermes was connected to a live geopolitical risk engine through a . The important design choice was to give Hermes facts about what happened and what changed, not predictions or advice. If asked whether to buy bitcoin, it refuses instead of inventing a signal.

Most of the work went into stopping Hermes from making up extra claims around real data. The same setup has run with a small , Claude, and DeepSeek. The model can change, while the trust rules stay attached to the data layer.

For now Hermes only observes, and action-taking will be considered after a 30-day trial.

Key points

  • Give Hermes observation data first, not advice or predictions.
  • Use a so Hermes can see data but cannot change or execute anything.
  • Block questions that would push Hermes to invent signals, such as direct advice.
  • Keep trust rules in the data layer so the setup can survive a model change.
  • Test observation-only behavior before adding tools that let Hermes act.
Read original