Skip to main content
Solana AI iconSolana AISAI Console
Policies

Operational guardrails for Solana AI.

Configure approval tiers, signer rotation schedules, and incident response plans so your team stays in control even as agents handle execution.

Tier 0 · Informational

Low-value or read-only operations that can run with minimal oversight.

  • Auto-approve requests below configured value thresholds.
  • Log actions to history with annotations for transparency.
  • Optional alerts to keep watchers informed without interrupting execution.

Tier 1 · Standard

Routine operations such as token transfers, liquidity rebalances, or strategy updates.

  • Require MPC quorum or two human approvers before execution.
  • Capture Switchboard simulation output when available.
  • Generate exports for finance teams after completion.

Tier 2 · Elevated

High-value transactions, signer rotations, or cross-chain movements.

  • Mandatory human approval plus MPC quorum threshold.
  • Out-of-band confirmation via Slack or PagerDuty.
  • Post-execution review within 24 hours, with findings logged to the history trail.

Tier 3 · Critical

Emergency actions such as halting strategies, draining treasuries, or revoking authorities.

  • Freeze the instruction queue until a designated incident team approves.
  • Run redundant simulations and require executive sign-off.
  • Document actions with a full incident report exported to compliance systems.

Operational procedures

  • Rotate signer keys on a configurable cadence or whenever an incident is detected.
  • Audit wallet sessions weekly—revoke stale sessions and clean up unused keys.
  • Review alert fatigue metrics to ensure operators respond quickly to critical events.
  • Keep your Neon MCP credentials scoped to least privilege and rotate secrets regularly.