SEC examination staff applying Advisers Act Rule 206(4)-7 expect an adviser using an automated platform to hold written policies assessing whether “algorithms were performing as intended”, supported by exception reports — and the same alert flags advisers whose white-label platform provider does not address this. This page is that artifact, generated from the chain rather than from any database we control.
POLICY SIGN-OFF · DORA ART. 5(2)(a)signed off on-chain · 2/2
Treasury Policy v1is the written mandate the accountable management body approves. Its canonical hash is a decision the independent auditor quorum votes on, so the approval isn't a claim — it's an on-chain fact. The agent embeds this policy version in every attested decision.
Escalate below confidence 75%Max rebalance 8% / moveMin reputation 1— live from the on-chain PolicyEngine, changes time-locked 60s (B4 governance)
Each row is a transaction a control refused. A refusal is not a failure — it is the control doing its job, and it is the evidence an examiner asks for.
Listed separately and deliberately: these are runtime/deployment conditions, not policy controls firing. Counting them as control refusals would overstate how often the guard rails engaged.
Testnet deployment — figures are demonstration values, not client assets.
Self-produced artifact. It evidences on-chain control behaviour; it is not an audit, an attestation engagement, or legal advice.
The zero-knowledge circuit has not been independently audited, so the solvency proof is not auditor-grade evidence yet.
Conduct and custody duties (MiCA/MiFID II) and ICT responsibility (DORA Art. 5(2)(a)) attach to the authorised provider — these controls make oversight provable, they do not transfer liability.