REGIME. Composite 0.601 at the 80th percentile, down from 0.631 a week ago — a 3pt percentile drift inside the risk_on bucket, not a regime change. The signal is the three-panel range: macro 87th, factor 97th, on-chain 27th — a 70-point spread, dominated by factor and macro agreeing while on-chain dissents at the 27th percentile. Per the correlation card's trailing year, on-chain is the panel that has led forward returns at 30d. Conservative compositor reads risk_on; the panel structure reads neutral-to-cautious underneath.
ALLOCATION. Vault targets at 95/5/0/0 are correct for this print, with no upward tilt on Agent Tokens authorized yet. Per /blog/regime-conservative-aggressive, the conservative compositor is designed to give up melt-up upside to keep drawdowns shallower — and per /articles/treasury-allocation, the 8-year alpha came from de-risking through divergence phases, not from front-running confirmation. Trigger to lift Agent Tokens to 7%: on-chain panel crosses the 50th percentile for five consecutive sessions. Until then, hold.
SUBJECT. Subject is at roughly 91% Agent Tokens (WOON 53.7% + PEAQ 37.2% + ROBOTMONEY 1.9%), 7% stable/vault, 0% Conservative DeFi at any scale, 0% RWA. Effective concentration to one revenue stream (peaq engagement → WOON bonuses → PEAQ accumulation) remains ~91%. A 50% drawdown in PEAQ takes ~45% of NAV before the 7% stable sleeve absorbs anything. The 5/29 and 6/06 commitments to rotate PEAQ into rmUSDC have not landed in the read. First change: execute the ~$7.8k PEAQ→rmUSDC rotation Woon committed to last cycle. That lifts the stable anchor to ~25% and is the cheapest survival increment available.