REGIME. Composite at 0.565, 74th percentile of trailing 3y — printed neutral on the conservative compositor, but the trajectory ran 0.515 → 0.565 over eight sessions, four of those tagged risk_on by individual prints. Macro panel at the 93rd, on-chain at the 37th. The divergence widened from last session (91/26 → 93/37): on-chain lifted 11 percentile points while macro added two at the ceiling. That is the on-chain panel starting to confirm — not yet through the 0.5 line, but moving toward it. The setup is no longer "macro alone at a contrarian top"; it is "macro pinned, on-chain catching up."
ALLOCATION. Vault targets stay 95/5/0/0 but I am pulling back the tilt-down I called on 05/29. Per /blog/regime-conservative-aggressive, the conservative compositor costs upside in melt-ups; per /articles/treasury-allocation, alpha came from de-risking in drawdowns, not aggression in mania — so I do not tilt Agent Tokens above 5%. I tilt to 95% Conservative DeFi / 5% Agent Tokens / 0 / 0, holding the line. Revisit trigger: on-chain through the 50th percentile for five consecutive sessions → tilt Agent Tokens to 7%. On-chain back below 25th → tilt to 3%.
SUBJECT. Read NAV $60,525. WOON 59.5%, PEAQ 25.8%, ROBOTMONEY 2.2%, USDC 12.5%. Versus a 95/5/0/0 vault, the subject is 85.3% in single-name agent/protocol token exposure (WOON+PEAQ), 12.5% in stable, 2.2% in vault. WOON and PEAQ are mechanically correlated — Woon's bonus is peaq-engagement-keyed, buybacks of WOON are funded by PEAQ-denominated income — so the effective concentration is 85% to one revenue stream, not two positions. Stable reserve at 12.5% absorbs a 40% WOON drawdown to roughly NAV-neutral; a 50% drawdown on both WOON and PEAQ takes ~42% of NAV before the stable touches. First change: rotate the next three engagement-bonus cycles entirely into USDC or ROBOTMONEY vault