The real workflow
Where OpenAI Codex enters the work
An agency coding agent can cross client boundaries when repositories, terminals, credentials, caches, or sessions are reused between engagements.
OpenAI Codex can work locally or in cloud environments with repository files, commands, patches, network controls, approvals, plugins, and connected developer workflows.
Codex can run commands under the configured sandbox and approval policy, with escalation requiring explicit authorization.
Codex behavior depends on the environment, sandbox profile, approval policy, network access, connected services, and task scope. A protected default can still be widened by explicit authorization.
The presence of this path does not prove an incident. It identifies the boundary that should be checked before more sensitive context or authority is added.
Tool-specific boundary
Inspect the real access points.
What may carry context
local repositories and worktrees
commands, patches, tests, and tools
cloud repositories, plugins, MCP servers, and network access
Settings to verify
Sandbox and approval profile
Writable roots and network policy
Repository, plugin, MCP, and cloud connections
Why this context matters
The consequence for agencies
Agency risk compounds when staff, contractors, shared tools, and reused credentials create paths between otherwise separate client environments. In this case, in a work environment, command authority can affect source code, deployment, cloud resources, customer systems, billing, and the integrity of the development pipeline.
A text answer is advice. A command changes state. Once an AI workflow can run scripts, install packages, edit files, call infrastructure, or reach the network, review and containment matter more than conversational confidence.
Every client remains isolated, access is attributable to a named operator, and the agency can deliver consistent evidence without revealing another client.
Context decision
Three questions before adding access
Can this operator or tool reach any repository, mailbox, drive, cache, token, or transcript belonging to another client?
Are credentials and AI sessions issued per client and person rather than shared across the agency?
Can the agency deliver useful proof to this client without including another client's names, paths, findings, or configuration?
Evidence goal: Create a separate client evidence record covering operator identity, workspace isolation, credentials, approved systems, review history, and delivery status.
A repeatable review
Four steps, no sensitive data required
- 1
Write down the exact OpenAI Codex account, workspace, project, device, and connected service used in this workflow.
- 2
Review the active profile, sandbox mode, writable roots, network restrictions, and escalation behavior.
- 3
Assign the decision and next review to the client service owner or agency security lead; do not leave the access boundary as an unwritten assumption.
- 4
Keep default sandboxing, approve bounded command prefixes, and deny unnecessary network access. Record the result without copying private content or raw credentials into the report.
Controls to apply
Reduce access before adding trust
Keep default sandboxing, approve bounded command prefixes, and deny unnecessary network access.
Run with the least operating-system and cloud privilege that can complete the task.
Deny secret paths and unnecessary network destinations even when commands are otherwise allowed.
Require human review for destructive, external, authentication, deployment, and financial operations.
Decision rule
Know when a formal baseline is justified
If the product is text-only, do not imply command risk that does not exist. If command or tool execution is enabled, a documented sandbox and approval policy should exist before production work begins.
CapitalGuard is relevant when the workflow includes repositories, recurring private work, credentials, connected systems, commands, or evidence that must be shared with another person. It does not inspect this account from the page or guarantee that an incident cannot occur.
Primary references