Long-Running Agent Job with Claude Code: MCP Tool Boundary

A production playbook for long-running agent job in cross-industry operations using Claude Code: mcp tool boundary, run-scoped inputs, logs, typed results, and artifacts.

Audience: AI product teams running multi-minute workflows

The problem

AI product teams running multi-minute workflows need long-running agent job to run repeatedly against documents, repos, tool access, and output contracts. In cross-industry operations, the pain is not one good answer; it is repeatability, auditability, exception handling, and evidence that survives handoff.

Implementation path

Expose only the MCP tools needed for long-running agent job, validate tool arguments, keep credentials in the owning service, and log each call outside the sandbox.

Tradeoffs and failure modes

Narrow tool boundaries reduce agent flexibility, but make the integration reviewable and supportable. For long-running agent job, the practical test is whether a second run can be debugged, retried, and consumed by a product without reading the raw agent transcript.

Tool policy

tool: long-running-agent-job_lookup
agent: Claude Code
input_scope: /skill/.argo/inputs
credential_owner: broker
log_arguments: true
network_policy: allowlisted

Run this on Argo