Handle Validator Slashing¶

Purpose¶

Move a slashed or slash-risk validator into fail-closed handling, protect principal accounting, and prepare exit/settlement.

Procedure¶

1. Verify slashing evidence from independent consensus sources.
2. Update risk observation with slashed=true through updateRiskObservationFinalModel.
3. Confirm risk state enters stop-loss or appropriate unsafe state.
4. Cancel or let controller cancel pending claims.
5. Disable executor path if broad claim automation could misfire.
6. Request validator exit when controller policy permits.
7. Track final exit withdrawal receipt and settlement readiness.
8. Propose and finalize settlement only after evidence supports it.

Abort Conditions¶

Abort payout enabling if observation is not finalized, sources conflict, vault/pubkey mapping is unclear, or reserve coverage cannot be proven.

Evidence To Archive¶

Consensus proof, risk update tx, claim-state readback, exit request, receipt records, settlement proposal/finalization txs, and communication timeline.

Operational Procedure¶

Purpose¶

Use this runbook to respond to slashing evidence and prevent unsafe claims or deposits from continuing.

When To Use¶

Use it for production, staging, or rehearsal actions that affect live authority, validator custody, economic accounting, claims, or incident response. Do not use it as a substitute for source review; deployed-state evidence remains Evidence required unless captured for the exact chain and address set.

Required Authority¶

Required authority: controller/risk owner with governance escalation. Read-only preparation can be performed by an operator or auditor, but transaction submission must come from the documented production holder in the permission matrix.

Preconditions¶

• The current source manifest and generated inventory are up to date.
• The acting Safe or owner has been verified against the current permission matrix.
• No unresolved incident is active for the same contract, validator, role, or operation.
• The reviewer can identify which layer is affected: Upgrade governance, Deposit permissioning, Custody/readiness, or Economic/claim safety.

Inputs Required¶

• v.
• a.
• l.
• i.
• d.
• a.
• t.
• o.
• r.
• .
• p.
• u.
• b.
• k.
• e.
• y.
• ,.
• .
• s.
• l.
• a.
• s.
• h.
• i.
• n.
• g.
• .
• p.
• r.
• o.
• o.
• f.
• ,.
• .
• e.
• p.
• o.
• c.
• h.
• ,.
• .
• v.
• a.
• u.
• l.
• t.
• ,.
• .
• l.
• a.
• t.
• e.
• s.
• t.
• .
• r.
• i.
• s.
• k.
• .
• o.
• b.
• s.
• e.
• r.
• v.
• a.
• t.
• i.
• o.
• n.
• ,.
• .
• r.
• e.
• s.
• e.
• r.
• v.
• e.
• .
• s.
• t.
• a.
• t.
• e.

Step-By-Step Procedure¶

1. verify slashing evidence.
2. submit or update risk observation.
3. cancel pending claims if policy does not already do so.
4. block new admission for the validator.
5. open settlement/exit assessment.

Independent Review Requirement¶

A second reviewer must check the decoded calldata, expected state transition, affected role or validator, and expected events before submission. For emergency use, capture the reviewer identity and incident ticket before or immediately after the transaction.

Abort Conditions¶

• Source manifest hash drift or unexpected implementation metadata.
• Caller or Safe does not match the permission matrix.
• Revert reason points at a different layer than the runbook is trying to change.
• Any required input is missing or only inferred.
• A guardian, canceller, or incident commander has frozen the action window.

On-Chain Pre-Checks¶

Read current role/owner/admin state, operation status, target code hash where applicable, validator/vault mapping where applicable, and the latest readiness or claim state that the action depends on. Record block number and RPC endpoint.

On-Chain Post-Checks¶

Confirm the intended state changed, no adjacent authority changed unexpectedly, and no pending operation or stale intent was left active. Re-read the affected contract rather than relying only on transaction success.

Events Or Logs To Monitor¶

• RiskObservationAccepted.
• ClaimCancelled.
• ExitRequested.

Evidence To Archive¶

Archive calldata, transaction hash, decoded event logs, pre/post reads, reviewer approval, incident or change ticket, and any source-manifest or release-artifact references used to justify the action.

Escalation Path¶

Escalate to governance signers for authority or upgrade anomalies, to controller/risk owners for economic or claim anomalies, to admission operators for intent mistakes, and to security incident response for unexpected code, role, or event drift.