Skip to content

Agent Integration Overview ​

AxonX gives agents discoverable Task contracts, execution interfaces, and research records. Use an existing agent host to operate AxonX, or use built-in sessions in Studio. Both paths share research services and workspaces; different parties manage model configuration and sessions.

Independent Agent integration paths

Choose a path ​

External agentBuilt-in agent
Entry pointHosts such as Codex or Claude CodeStudio → Agent, or the agent_chat Job
Model configurationManaged by the external hostServer-side Claude Agent SDK configuration
AxonX connectionDirect CLI access or service /mcpBuilt-in component invokes configured Jobs through Dispatcher
InstructionsAxonX Skill and development guideOptionally load the development guide bundled with the package
Tool scopeAvailable CLI commands or public service MCP catalogjob_tools, SDK tools, and permission settings
Session recordsManaged by the external hostManaged by AxonX session storage
Reading pathExternal agents → MCP integrationBuilt-in configuration → Built-in usage

External integration does not require CLAUDE_CODE_* settings or built-in guide loading. The built-in agent's eight default Job tools query tasks and files; the public external MCP catalog may also include submission, cancellation, and installation. Actual capabilities depend on the selected path's configuration and schemas.

The shared research loop ​

  1. Establish the service address, credentials, and workspace; query machine resources and installed plugins.
  2. Discover Task definitions and read their actual input and output schemas.
  3. Reuse successful upstream tasks and execute only the stages the experiment needs.
  4. Keep the returned task_id and run_id, and wait for that execution's terminal state.
  5. Inspect logs, metadata, artifacts, and lineage before forming a research conclusion.

Keep the same --target across stages when using direct CLI access. Studio forwards requests through its same-origin backend to the selected machine. Tasks, data, and plugins live in the execution service's environment; agent integration does not migrate them automatically.

Continue by goal ​

Agent-native quant research.