optimize-ousd-allocation

Analyzes OUSD collateral across Morpho strategies and recommends yield-maximizing rebalancing plans.

152|115|Updated Aug 3, 2020
One-click install
npx skills add https://github.com/OriginProtocol/origin-dollar --skill optimize-ousd-allocation-originprotocol
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: optimize-ousd-allocation
Source: https://github.com/OriginProtocol/origin-dollar/tree/main/.claude/skills/optimize-ousd-allocation
Command: npx skills add https://github.com/OriginProtocol/origin-dollar --skill optimize-ousd-allocation-originprotocol

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Deciding how to allocate OUSD collateral across Ethereum, Base, and HyperEVM Morpho strategies, the Curve AMO, and the vault buffer requires reconciling live on-chain state, transient versus sustainable APYs, and production constraints — a manual process prone to trusting stale or misleading data. ## Core Features & Use Cases - Independent on-chain verification: Reads deployed balances, utilization, and liquidity directly via cast rather than trusting the MCP or the production rebalancer feed. - Sustainable-rate optimization: Values each market at its 90%-utilization rate and applies production gates (0.5% spread floor, $5K/$25K minimums, $50K steps, freeze thresholds) to produce an executable net-moves plan. - Learning loop: Logs every decision, scores past recommendations against realized Grafana history, and accumulates market-sensitivity heuristics in LEARNINGS.md. - Use Case: Ask whether to move funds from Base to HyperEVM and receive a sized, constraint-checked plan with expected $/yr delta, pre-flighted against on-chain liquidity. ## Quick Start Ask the assistant to optimize the OUSD allocation and recommend any rebalancing moves across the Morpho strategies, AMO, and vault buffer.

Frequently Asked Questions about optimize-ousd-allocation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I optimize OUSD allocation across Morpho strategies?

Run the skill to fetch live balances and APYs from the MCP and on-chain, then it greedily allocates the movable budget in $50K steps, valuing each market at its 90%-utilization sustainable rate. It outputs net moves with expected annual dollar gains but never executes transactions.

Why does the skill not trust the MCP or production rebalancer data?

The MCP's maxPossibleAmount is a withdrawal ceiling, not deployed balance, and it cannot show cross-chain utilization. The production rebalancer feed has been observed stale or divergent from on-chain state, so the skill reads checkBalance and Morpho Blue markets directly with cast and treats external feeds as cross-checks only.

What constraints govern OUSD rebalancing recommendations?

Moves must clear a 0.5% APY spread, $5K minimum ($25K cross-chain), and $50K step granularity. Strategies above 50% APY are frozen, the AMO is held at a 10% TVL floor, withdrawals may not push markets above 90% utilization, and the vault withdrawal shortfall is always funded first.

Can the skill execute rebalancing transactions on-chain?

No. The skill produces a plan only — net moves with expected annual yield delta and the executing call per leg. Actual relocation is a Strategist-multisig action performed by a human after reviewing the plan.

How does the skill handle thin cross-chain markets like HyperEVM?

It discounts headline APYs when on-chain utilization exceeds 90%, since those rates are transient spikes that revert. Cross-chain moves are sized small and staged with re-probing between tranches, because thin markets can move hundreds of basis points per $50K.

How are past rebalancing decisions reviewed and scored?

Each plan is appended to decisions/decision-log.jsonl with the projected sustainable destination rate. A later review pass queries Grafana morpho_vault_apy history over the following 7 days, compares realized APY to projections, assigns a good/marginal/bad verdict, and feeds calibration deltas into LEARNINGS.md.