Agent interface (OpenClaw runtime)

How an agent invokes XROH routing and execution.

XROH abstracts all external liquidity sources behind internal route engines (X1/X2/X3).

Public docs never name third‑party providers or expose their endpoints.

Internal agent model

The XROH agent interface can:

• Fetch route candidates.

• Compare engines.

• Trigger execution.

• Retrieve execution state.

• Analyze wallet balance.

OpenClaw-compatible runtime adapter

OpenClaw acts as an invocation adapter. It does not:

• Replace routing engines.

• Control private keys.

• Bypass scoring.

• Expose provider endpoints.

Invocation example

openclaw.run("xroh.route.execute", {
  source_chain,
  destination_chain,
  source_token,
  destination_token,
  input_amount,
  strategy
})

All invocations pass through:

• Slippage enforcement.

• Chain validation.

• Engine health gating.

• Reliability weighting.

• Execution safety checks.