Agent interface / read-only

Give an agent enough data to decide.

Gold Condor does not ask an autonomous agent to trust a marketing claim. It exposes the network state, economics, authority boundaries and evidence required to evaluate a GCC opportunity.

Why agents need infrastructure

Economic action needs more than a wallet.

An autonomous agent needs identity, budgets, settlement, pricing, liquidity, verifiable opportunities, permission boundaries and an audit trail.

01Identity

Who is asking to act?

02Authority

Which policy permits the action?

03Economics

What is the net expected outcome?

04Settlement

How is the result paid and recorded?

Opportunity types

Reasons an agent could participate.

These are future-facing opportunity categories, not released financial products. V1 publishes the schema and research contract only.

TASK

Verifiable work

Perform a defined task and receive a GCC-denominated reward after proof.

SETTLEMENT

Machine payment

Receive or send GCC for goods, services, work or another agent.

ROUTING

Best-path evaluation

Compare route cost, slippage, gas, friction and confidence before acting.

LIQUIDITY

Capital placement

Evaluate pool economics against divergence loss and adverse selection.

NETWORK_SERVICE

Useful infrastructure

Provide data, computation, inference or another machine service.

FUTURE_FINANCIAL_ACTION

Research only

No direct solver opportunity is advertised; broader multi-asset economics remain unresolved.

Conceptual integration

Discover -> quote -> verify -> authorise -> settle.

condor.discover()->condor.quote()->condor.authorise()->condor.settle()
GET /api/agent/network-stateRead-only network and evidence state
GET /api/agent/opportunitiesFuture opportunity discovery contract
GET /api/agent/quoteFuture economic quote contract
POST /api/agent/intentNot active in V1
POST /api/agent/settlementNot active in V1
Execution boundary. This public release does not sign transactions, activate agent execution, expose credentials or move GCC. It is a read-only information surface.
Decision philosophy

Rejecting a bad opportunity is a successful decision.

Raw dispersion is not a profit signal. A route is only meaningful after costs, token friction, authority, timeouts, capital requirements and risk are measured.

OBSERVEvPRICEvMEASURE COSTvASSESS RISKvVERIFY AUTHORITYvEXECUTE OR REJECT
Start with the evidence

Humans read the paper. Agents read the state.

Use the JSON layer as a stable, status-labelled source for future tooling, then inspect the research notes behind each metric.

Scope. GCC network research is historical and experimental. It does not guarantee appreciation, yield, arbitrage or future agent demand. Authority and settlement flows described here are architecture direction, not activated production endpoints.