Banking: Suspicious Activity Review Packet with Codex: Sandbox Policy
A production playbook for suspicious activity review packet in banking operations using Codex: sandbox policy, run-scoped inputs, logs, typed results, and artifacts.
Audience: Financial crime teams
The problem
Financial crime teams need suspicious activity review packet to run repeatedly against transaction exports, customer profiles, alerts, case notes, and SAR criteria. In banking operations, the pain is not one good answer; it is repeatability, auditability, exception handling, and evidence that survives handoff.
Implementation path
Run suspicious activity review packet 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 suspicious activity review packet, 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: Codex
Run this on Argo