Artifact Delivery with Claude Code: MCP Tool Boundary
A production playbook for artifact delivery in cross-industry operations using Claude Code: mcp tool boundary, run-scoped inputs, logs, typed results, and artifacts.
Audience: Teams whose agents create files customers download
The problem
Teams whose agents create files customers download need artifact delivery to run repeatedly against reports, exports, patches, PDFs, and manifests. 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 artifact delivery, 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 artifact delivery, 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: artifact-delivery_lookup
agent: Claude Code
input_scope: /skill/.argo/inputs
credential_owner: broker
log_arguments: true
network_policy: allowlisted
Run this on Argo