Skip to main content
Veda-style guardrails are the reference model for future AlphaEngine capital deployment. The key idea is simple: capital can follow selected strategies only inside a bounded vault policy.

Policy before deployment

A strategy cannot bypass vault policy because it ranked well.

Evidence feeds review

Arena evidence informs eligibility and pilot sizing.

Monitoring continues

Guardrails remain active after selection.

​
Guardrail categories

CategoryExample constraints
Asset allowlistonly approved markets and tokens
Size limitsmax notional, max share of vault, max market exposure
Execution permissionsapproved adapters, routes, and operators
Risk limitsdrawdown, tail risk, turnover, concentration, liquidity
Operational boundspause, review, upgrade, and emergency procedures
Reportingartifact hashes, report hashes, final effects

​
How this connects to the arena

The arena produces evidence. Guardrails decide what that evidence can unlock.
1

Arena evidence

Simulation and evaluation artifacts establish the candidate record.
2

Eligibility review

Review checks evidence quality, gates, and policy fit.
3

Bounded pilot

Any future deployment should be size-limited and policy-constrained.
4

Continued monitoring

Guardrails, reporting, and pause conditions continue after selection.
The strongest candidates should still be limited by the vault policy. A strategy cannot bypass asset, size, or risk constraints because it ranked well in one epoch.

​
Documentation rule

When describing the capital path, write:
Evidence creates eligibility. It does not guarantee capital.
Do not imply automatic deployment or fixed returns.

Guarded vaults

Future capital path under explicit constraints.

Scoring and benchmarks

Score evidence that can feed review.

Public settlement

What becomes observable after execution.