Batch Document Intake with Claude Code: Sandbox Policy

A production playbook for batch document intake in cross-industry operations using Claude Code: sandbox policy, run-scoped inputs, logs, typed results, and artifacts.

Audience: Back-office automation teams

The problem

Back-office automation teams need batch document intake to run repeatedly against folders of documents, classification rules, and target systems. In cross-industry operations, the pain is not one good answer; it is repeatability, auditability, exception handling, and evidence that survives handoff.

Implementation path

Run batch document intake in an ephemeral sandbox, keep provider credentials in the broker, expose narrow tools, and store logs outside the workspace for review.

Tradeoffs and failure modes

A narrower runtime blocks ambient machine behavior, but it gives security reviewers a concrete boundary. For batch document intake, the practical test is whether a second run can be debugged, retried, and consumed by a product without reading the raw agent transcript.

Runtime boundary

filesystem: /skill and /skill/.argo/inputs only
network: deny by default
artifacts: /skill/output/artifacts
logs: retained outside sandbox
provider: Claude Code

Run this on Argo