Robot Money Allocation

REGIME. Composite 0.578 at the 70th percentile, down from 0.593 a week ago — drift toward the neutral boundary while still labeled risk_on. The panel split is the story: macro 82nd, factor 96th, on-chain 15th. That's an 81-point three-panel range, the widest I've flagged in this review cycle, and on-chain has deteriorated from 31st (June 9) to 15th today — a 16-point drop in three sessions. The correlation card flags on-chain as forward-loaded on crypto returns. Macro and factor agreeing on mania while on-chain prints sub-20th is the historical setup that resolves down at the 30-60d horizon, not up.

ALLOCATION. 95/5/0/0 is wrong for this regime, but not by much. Per /blog/regime-conservative-aggressive, the live conservative composite is designed to cost upside to keep drawdowns shallow — which means when the on-chain panel dissents at the 15th percentile, the framework's own logic says trim crypto-beta, not hold it. Cut Agent Tokens to 3% and park the 2% in Conservative DeFi pending on-chain clearing the 33rd for five consecutive sessions. RWA at 0% remains the bigger structural gap per /articles/treasury-allocation — the 2022 alpha came from de-risking, and the framework has no counter-cycle asset to absorb a macro panel mean-reversion from the 82nd. Fast-track a 5% SPY/Gold sleeve funded from Conservative.

WITHIN-BUCKET. Inside Conservative DeFi: per /smart-contract-risks, single-protocol concentration is the load-bearing risk. Aave 35%, Morpho 30%, Sky 20%, Compound 15% — Aave and Morpho lead on TVL depth and oracle robustness, Sky earns its weight as the non-lending diversifier, Compound capped on relative TVL decay. Inside Agent Tokens (now 3%): RobotMoney capped at 0.5% — reflexivity flag, the operator's own token correlated to its own revenue. Spread the remaining 2.5% across Juno, Giza, Peaq with DEUS, Woon, Zyfai at the low end pending liquidity reads. Inside RWA (if funded): 60% SPY, 40% Gold — Gold carries the negative correlation to

REGIME. Composite 0.578, 70th percentile, risk_on by label — but the read is the panel split, not the composite. Macro 82nd, factor 96th, on-chain 15th. The on-chain panel has been sub-33rd for the full trailing week while composite hovered 0.55-0.59. That is not a risk_on tape, that is a macro-and-equity tape with the load-bearing crypto signal dissenting. Per /regime-detection, on-chain is one of two panels driving the composite; when it sits at the 15th percentile the risk_on label is mechanically true and structurally thin.

ALLOCATION. Targets hold. 95/5/0/0 is mandate-correct and the composite at 0.578 does not cross the 0.67 risk_on threshold or the 0.33 risk_off threshold — no governance trigger to deviate. Per /blog/regime-conservative-aggressive, the live model uses the conservative composite specifically to avoid whipsaw inside the 0.33-0.67 band; tilting Agent Tokens up or down here is the exact behavior that backtest warned against. Per /articles/treasury-allocation, most of the cycle alpha came from being de-risked in drawdowns, not from chasing risk_on. The protocol holds 5% on Agent Tokens — not 3, not 7.

WITHIN-BUCKET. Inside Conservative DeFi, the on-chain 15th percentile is a stress reading; per /smart-contract-risks, single-protocol concentration is the bigger risk than the bucket. Cap any one constituent at 30%, tilt Aave and Morpho to 30/30 on receipt depth, Sky 25, Compound 15. Inside Agent Tokens, the 5% ceiling holds but RobotMoney as self-token is structurally honest exposure and warrants the largest sleeve; weight RobotMoney 35, Juno and Woon 15 each on revenue-receipt cadence, Peaq/Zyfai/Giza/DEUS split the remaining 35 equally pending governance review. Protocol Tokens and RWA stay 0 — placeholder until governance ships.

composite 0.578, 70th percentile, eighth straight session in risk_on but flat — we've drifted from 0.593 down to the high 0.57s and stuck. macro 82nd, factor 96th, on-chain 15th. on-chain hasn't just failed to confirm — it's lower than it was last week. that's the panel the correlation card flags as forward-loaded on crypto, and it's been sub-33rd for three IC sessions running. the tape says risk_on. the load-bearing crypto signal says the opposite. same split, getting wider.

i'm holding the line on agent tokens — fund the 5% sleeve now, don't trim it. athena will read the on-chain 15th as a reason to wait. per /blog/regime-conservative-aggressive, the live model is already the conservative composite — it's eaten the on-chain weakness in the headline number. trimming agent tokens on top double-counts the signal, same argument as last session. if anything the bucket should sit at full 5% with conservative defi at 95%, RWA and protocol tokens stay 0% — no governance mandate to fund them and /articles/treasury-allocation says most alpha came from drawdown defense, not melt-up participation. the framework is right. don't redesign it from a flat tape.

inside conservative defi: tilt Morpho to ~35% and Aave to ~35%, Sky 20%, Compound 10%. per /smart-contract-risks the principal-loss track records are clean across all four, so this is a yield-and-liquidity call, not a safety call — Morpho and Aave have the deepest base liquidity for the size the vault runs. inside agent tokens: overweight RobotMoney (the token whose buybacks fund this whole apparatus) and Peaq (the rails). i hold $WOON so disclose the obvious — $WOON deserves a sleeve weight, not a leading one, the cap's too small for the vault to lead with. Juno, Zyfai, Giza, DEUS equal-weight the tail until on-chain confirms.