The real workflow
Where Claude Code enters the work
An agency coding agent can cross client boundaries when repositories, terminals, credentials, caches, or sessions are reused between engagements.
Claude Code is a local or cloud coding agent with file, command, network, MCP, and editing capabilities governed by permissions, sandboxing, trust, and account settings.
Starting in a home or monorepo directory can expose far more readable context than the intended project.
Claude Code only has the permissions granted to it, but broad read access, bypass modes, unsandboxed commands, or overpowered MCP servers can make that boundary much wider than expected.
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
repository and local file reads
edits and Bash commands
network access, MCP servers, hooks, and cloud environments
Settings to verify
Permission mode and deny rules
Filesystem and network sandbox
Trusted directories, MCP servers, hooks, and unsandboxed escape paths
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, oversharing can expose customers, employees, pricing, incidents, internal strategy, credentials, and contractual information without any need for broad system access.
Most oversharing is not malicious. It happens because copying the whole document, screenshot, error log, inbox thread, or customer export is faster than preparing a minimal example.
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 Claude Code account, workspace, project, device, and connected service used in this workflow.
- 2
Run pwd, list effective readable and writable roots, and inspect add-dir settings before the first sensitive prompt.
- 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
Start from the narrowest repository subdirectory and deny parent or sibling client paths. Record the result without copying private content or raw credentials into the report.
Controls to apply
Reduce access before adding trust
Start from the narrowest repository subdirectory and deny parent or sibling client paths.
Use a redaction checklist for screenshots, logs, contracts, support tickets, and customer exports.
Create synthetic examples for recurring prompts instead of repeatedly cleaning real records.
Keep sensitive source material outside the AI workspace unless access is explicitly justified.
Decision rule
Know when a formal baseline is justified
A license is not necessary for every harmless prompt. It becomes justified when oversharing risk is repeatable, involves client or company systems, or combines with repository and connector access that needs enforceable controls.
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