Skip to main content

Training / Operational Mode

BASTION implements a global mode toggle that separates training exercises from operational use, following the "train as you fight" design principle — the tools and workflows are identical in both modes.

Global Toggle​

A top-level application toggle switches between:

  • Training Mode — for exercises, wargames, and course of action analysis.
  • Operational Mode — for real-world planning and execution.

The toggle is set at the application level and affects all users in the current session.

Training Mode​

When training mode is active:

  • An amber EXERCISE banner is displayed at the top of the application.
  • EXERCISE watermarks appear on all exported documents and screenshots.
  • All data is written to a training-isolated context — no contamination of operational data.
  • AI agents operate against exercise scenario data only.

These visual indicators ensure participants always know they are working in a training environment, preventing accidental confusion with operational systems.

Operational Mode​

When operational mode is active:

  • The UI is clean — no banners or watermarks.
  • Data is written to the operational context.
  • AI agents operate against real-world data and intelligence products.

Governance Parity​

DAO governance works identically in both modes:

  • Proposals, voting, quorum rules, and role-weighted decisions follow the same process.
  • Training exercises practice the full governance workflow so operators are proficient when it matters.
  • This parity is intentional — governance shortcuts in training would create bad habits.

Reset and Checkpoint​

Training mode supports exercise iteration:

  • Checkpoint — save the current exercise state (database snapshot) at any point.
  • Reset — restore to a previous checkpoint to re-run a phase or decision point.
  • Full reset — return to the initial scenario seed state.

This allows facilitators to replay critical decision points, explore alternative courses of action, or restart after a teaching moment.

After-Action Review​

Training exercises capture data for post-exercise analysis:

  • Decision logs with timestamps and rationale.
  • AI agent recommendations and whether they were accepted or overridden.
  • Planning timeline and phase gate completion metrics.
  • Staff role participation and collaboration patterns.

This data supports structured after-action reviews (AARs) and feeds into future exercise design.