Obsqra.fi Docs
Demo

DAO-Constrained Allocation

DAOs define hard limits on AI behavior through on-chain policies. Recommendations that satisfy all constraints can auto-execute; others require governance approval.

The Governance Challenge

DAOs face a dilemma when delegating capital management to AI:

Too Much Trust

Too Much Trust

Giving AI full autonomy means the DAO loses control. A rogue or manipulated AI could drain the treasury into a single high-risk protocol.

Too Little Trust

Too Little Trust

Requiring governance votes for every rebalance is slow and expensive. Markets move faster than DAOs can vote.

Obsqra.fi solves this with constraint-based autonomy: the DAO defines the boundaries, and the AI operates freely within them.

How Constraints Work

DAO Vote

Set policy parameters

AI Recommends

Generate allocation

Check Constraints

On-chain validation

Auto-Execute

If all constraints pass

Constraint Types

Max Single Protocol Allocation

Prevents concentration risk by limiting how much can go to any single protocol.

Typical:5000(50%)
Conservative:3000(30%)
Minimum Diversification

Requires allocation across multiple protocols to reduce single-point-of-failure risk.

Typical:3protocols
Conservative:5protocols
Max Volatility Score

Sets a ceiling on the average risk score across the portfolio.

Typical:75/100
Conservative:50/100
Max Allocation Change

Limits how much the allocation can shift in a single rebalance, preventing sudden large movements.

Typical:2000(20% shift)
Conservative:1000(10% shift)

Auto-Approval Flow

When the AI generates a recommendation, the contract checks each constraint.

Scenario
40/35/25 split, risk score 45
Auto-Execute?
Yes
Reason
All constraints satisfied
Scenario
60/25/15 split, risk score 45
Auto-Execute?
No
Reason
Exceeds max single protocol (60% > 50%)
Scenario
50/50/0 split, risk score 45
Auto-Execute?
No
Reason
Only 2 protocols (< 3 minimum)
Scenario
40/35/25 split, risk score 80
Auto-Execute?
No
Reason
Volatility too high (80 > 75)
Scenario
Manual approval enabled
Auto-Execute?
No
Reason
DAO requires vote for all changes

Governance Override

When auto-execution is blocked, the recommendation enters a governance queue.

Constraint Failed

Auto-execute blocked

Queue

Enters governance

DAO Vote

Token holders decide

Execute

If approved

ℹ️
Emergency Override
The DAO can also set requiresManualApproval = true during market stress to pause all auto-execution. This is a safety mechanism that does not require changing individual constraint values.

Best Practices

Do

  • +Start with conservative limits
  • +Gradually relax as trust builds
  • +Monitor constraint violations
  • +Have emergency pause capability
  • +Review policies quarterly

Don't

  • -Set limits too loose initially
  • -Ignore constraint violations
  • -Remove all constraints
  • -Use single-sig for governance
  • -Forget to test edge cases