Threat Model¶

Assets To Protect¶

• validator principal protection outcomes
• reward accounting correctness
• governance integrity
• admission and vault-binding integrity
• availability of claim and settlement operations

Trust Boundaries¶

• governance multisig and role custody
• oracle/backend submission authority
• deployment and release artifact integrity
• reserve proof integrity and freshness

Major Threat Classes¶

• governance key compromise
• malicious or mistaken implementation approval
• deposit baseline misconfiguration
• oracle replay/stale/conflict manipulation
• receipt source replay or misclassification
• operator runbook execution error

Expected Defensive Behaviors¶

• fail-closed when state evidence is stale or conflicting
• strict role checks on privileged actions
• idempotency checks for receipts and economic sources
• deterministic claim caps and pending-claim lifecycle controls

Residual Assumptions¶

• signers are not compromised
• offchain observation and receipt pipelines are honest enough for fail-closed model inputs
• governance process can detect malicious queued operations within timelock window