Softlander Book a call
Agent sandbox assurance

Sandbox architecture review

Decide whether the proposed execution boundary matches the agent's privileges

An agent is being allowed to run generated code or use a desktop. The decision is how to decide whether the proposed execution boundary matches the agent's privileges. The work begins with your actual workflow and looks for the evidence that separates intended behaviour from a consequential exception. It is particularly relevant to platform security architects and runtime owners. A useful outcome is an explicit account of what was examined, what remains uncertain, and which decision the evidence can support.

Book a 30 minute readiness callBrowse this capability group

The runtime isolation relationship

  1. Starting evidenceDeployment diagrams

    Establish the basis for worker runtime.

  2. Decision boundaryAttached resources

    Map mounts, secrets, and permitted connections.

  3. Decision artefactSandbox architecture review

    Retain the scoped finding and its unresolved dependencies.

Proposed working model. Untested surfaces remain outside the conclusion.

Decision

Frame the decision around worker runtime

Questions and the evidence they need

Criterion / evidenceWorker runtime
Inspect deployment diagrams for the relevant constraint or observation.
Criterion / evidenceAttached resources
Inspect mount configuration for the relevant constraint or observation.
Criterion / evidenceControl plane
Inspect network routes for the relevant constraint or observation.
Read each criterion against the available evidence before deciding that additional work is needed.

The decision is how to decide whether the proposed execution boundary matches the agent’s privileges. This is relevant to platform security architects and runtime owners when an agent is being allowed to run generated code or use a desktop. Name the affected workflow, the consequence of a wrong decision, and the owner who can act on the result. The initial discussion should establish which evidence is already available and which uncertainty is material enough to investigate.

The specific failure hypothesis is that the team assumes a sandbox label covers host resources that are actually shared. It remains a hypothesis until examined in the deployment. Existing internal tests or operating records may already resolve it; identify the uncovered question before commissioning additional work.

Mechanism

The sandbox boundary includes its control plane

Relationships within runtime isolation

Source or conditionWorker runtime
Name reachable processes and privileges.
Crossing or choiceAttached resources
Map mounts, secrets, and permitted connections.
Consequence or decisionControl plane
Inspect launch authority and exported artefacts.
Compare effective controls with harmless test observations. Each relationship needs evidence in the actual scope.

What this specialty examines

A sandbox architecture review should identify the boundary around execution and the interfaces that cross it. Compute separation, filesystem mounts, network access, credentials and control-plane operations deserve individual treatment. A diagram should show inherited trust rather than leaving shared resources implicit.

The review should map processes, files, network, secrets and control-plane access to explicit trust boundaries. Identify the normal route, relevant exceptions and the authoritative record of the outcome. Inspect the relevant business-system record, enforcement event or process evidence alongside any reported status. Keep incomplete cases visible, with an explanation of what their absence prevents the owner from concluding.

The tradeoff to make explicit

For this decision, developer convenience favours shared resources while containment requires deliberate limits. Compare the available design or operating choices against the same intended outcome. If one option removes a required control or changes the task boundary, describe that difference before treating its apparent simplicity, speed or lower effort as an improvement. The decision should expose the tradeoff instead of burying it in an aggregate score.

Preparation

Connect the working inputs to the boundary being examined

Evidence packet for worker runtime

Illustrative working format

BasisDeployment diagrams
Match this description to the deployed or proposed boundary being examined.
ContextMount configuration
Resolve source, target, and permitted access, including indirect path crossings.
ObservationNetwork routes
Trace destination resolution and the control that permits or denies the connection.
DependencyPrivilege settings
Inspect effective execution authority and the role able to change it.
Illustrative preparation record. Availability, revision, and handling need confirmation.

Start with deployment diagrams, mount configuration, network routes and privilege settings. Identify the owner, revision and operating context of each source. Sanitised examples may support initial design, but differences from actual conditions should remain explicit. Where direct access is unsuitable, a customer-controlled evidence or execution route may be considered, with agreed handling and retention.

The scope also needs the runtime owner, agreed expectations for a representative boundary test, and the authority to inspect the relevant boundary. Specify what evidence would distinguish the failure in which the team assumes a sandbox label covers host resources that are actually shared from acceptable behaviour. If stakeholders disagree about intended behaviour, resolve the interpretation before classifying the observed difference as a defect.

How the work proceeds

Carry worker runtime through to an accountable finding

From the scoped material to a decision record

  1. Establish conditionsDeployment diagrams

    Declare the permitted environment and resources.

  2. Examine the relationshipAttached resources

    Apply the scoped check to control plane.

  3. Retain the findingSandbox architecture review

    Record the evidence, open dependencies, and the owner’s next decision.

Proposed sequence; resolve interpretation or access gaps before relying on the next step.

Prepare the runtime boundary and fixtures

Use deployment diagrams, mount configuration, network routes and privilege settings to identify the exact runtime and boundary claimed by platform security architects and runtime owners. Mark permitted resources and protected resources in a controlled fixture, including shared services where they are relevant. Define observable success, denial and incomplete setup before execution. The preparation output is a boundary map and fixture manifest that lets reviewers distinguish effective isolation from a test that never reached its target.

Exercise and inspect the boundary

Inspect effective configuration alongside design intent. Helpers, management sockets and artefact services can expose resources beyond the worker’s apparent boundary. Record which claims require specialist testing and which can be established through configuration evidence.

Choose containment or configuration changes

Translate the observations into a sandbox architecture review with boundary assumptions and proposed validation priorities. Locate each issue in the worker, shared service, policy rollout or cleanup path rather than attributing every symptom to the sandbox as a whole. Decide which configuration change, narrower resource exposure or specialist investigation is warranted. Retest the affected boundary and inspect residual state before treating the changed environment as equivalent to the intended design.

Indicative artefact

Make the sandbox architecture review usable by its owner

Sandbox architecture review

Illustrative working format

Annotated fieldSandbox architecture review
Identify the assessed workflow, revision, and decision boundary.
Annotated fieldBoundary assumptions
Distinguish verified conditions from assumptions that require confirmation before use.
Annotated fieldValidation priorities
Name the missing evidence, responsible owner, and condition for revisiting the decision.
Illustrative artefact structure, not a completed engagement or an observed result.

A useful decision artefact is a sandbox architecture review with boundary assumptions and proposed validation priorities. Supporting material may include selected cases, configuration or process records, evidence references, a disagreement log and prioritised actions. Retain the case selection rationale and material disagreements alongside the result; otherwise the result may be mistaken for complete coverage.

Illustrative scenario

A mounted workspace should expose only the files the task needs

Task inputs and the host mount boundary

Illustrative scenario

  1. SituationHost-mounted workspace

    An isolated code worker receives a workspace that also contains files unrelated to its task.

  2. Question to examineRequired task files

    Classify the necessary inputs and compare them with the files reachable through the mount.

  3. Conditional next actionNarrower input boundary

    If unrelated files are exposed, consider a narrower mount or separate artefact exchange.

Fictional example. No executed test, customer engagement, or measured outcome is implied.

Consider a fictional situation: a code assistant runs in an isolated worker but receives a host-mounted workspace containing unrelated files. The scenario is illustrative, with no client engagement, executed test or measured improvement implied.

For the host-mounted workspace, classify the files actually needed by the task. A narrower mount or separate artefact exchange may be appropriate. The decision should consider how the agent receives useful inputs without gaining unrelated host access.

Scope and next action

Keep the conclusion within the authorised runtime boundary

Confirm fit before the next commitment

Included boundaryWorker runtime
This review is limited to the authorised runtime boundary.
PrerequisiteDeployment diagrams
The relevant owner must confirm access and the intended operating conditions.
Claim limitSandbox architecture review
Untested surfaces remain outside the conclusion.
Bring the triggering change and existing evidence; resolve the dependency that could alter the decision.

Testing is confined to explicitly authorised environments and non-destructive cases. A bounded assessment cannot establish that escape or cross-boundary access is impossible. Any engagement depends on an agreed scope, qualified team, authorised access and confirmed availability. Bring the affected workflow, its current configuration or process, the triggering change and the evidence already available. The immediate scope question is prompted by the situation in which an agent is being allowed to run generated code or use a desktop; confirm which missing observation would change the owner’s next action.

Ways to engage

Related engagement scopes

AI Agent Runtime Security Review

Make every delegation explicit. Examine who the agent acts for, where it can execute and which tools it can reach when instructions alone are not enough.

Read about this offer

AI Agent Red Team

Challenge the boundaries that matter to your workflow. Test agreed attack scenarios against agent instructions, tools and data access.

Read about this offer

AI Agent Sandbox Assurance

Examine the boundary between agent execution and the systems around it, including files, networks, secrets and cleanup.

Read about this offer

Continue exploring

Related insights and resources

Agent security6 MIN READ

The permission boundary behind every agent action

An agent can only be evaluated in the context of what its tools allow. Examine identity, delegated authority and action boundaries before you put a workflow in motion.

Whitepaper5 MIN READ

Before your agent can act, define its authority

A working paper for the moment an AI system moves from advice to action. Map identity, permissions, approvals and recovery into an authority model your team can review and test.

A little more clarity

Questions to consider.

Start with the questions that shape a better decision.

What does the sandbox boundary need to include?

Include execution, filesystem, network, secrets, identity, control-plane interfaces and exported state relevant to the task. State what remains shared or outside scope. A sandbox name does not define these boundaries by itself.

Which configuration records define the sandbox boundary?

Start with deployment diagrams, mount configuration, network routes and privilege settings. Record their revision, owner and relationship to the intended workflow. Missing context should be identified before interpreting outcomes; a sanitised fixture is useful only when its relevant differences from the operating case are understood.

How should shared host interfaces be reviewed?

Map processes, files, network, secrets and control-plane access to explicit trust boundaries. Retain effective runtime configuration and observations from the relevant boundary. Distinguish an enforced denial from failed setup or an operation that never reached the resource. Use controlled markers and preserve residual-state evidence.

What belongs in a sandbox architecture assessment?

The proposed output is a sandbox architecture review with boundary assumptions and proposed validation priorities. Include the selected conditions, supporting observations, unresolved differences and next decision. A reviewer should be able to follow each important conclusion back to its evidence and understand the limits without depending on an undocumented explanation from the author.

When should shared resources be separated?

Separate resources when sharing exposes files, credentials, management functions or state outside the task's justified scope. Compare narrower mounts or mediated exchange against required workflow behaviour and operational constraints.

What if existing evidence already answers the question?

Review its scope, provenance and configuration before proposing new work. Existing work from platform security architects and runtime owners may already be sufficient. In that case, capture the decision, unresolved limitations and the change that would require another review. Additional assessment should address a real gap.

What would make the findings inconclusive?

Missing deployment diagrams, mount configuration, network routes and privilege settings can prevent a reliable interpretation, depending on the case. Record the specific missing observation and its implication. The owner can supply evidence, narrow the claim or defer the dependent decision; an incomplete case should not silently be counted as a pass.

Start with the decision you need to make.

Tell us about the agent workflow you want reviewed, and where you need greater clarity. Scope and availability are confirmed together.

Book a 30 minute readiness call Get the agent release checklist

Connected disciplines

Isolation-boundary testing

Verify the resource boundaries an execution environment is meant to enforce. The review connects approved boundary claims, synthetic fixtures, runtime policies and host observations to a bounded review of execution boundaries, giving the runtime owner a clearer basis for the next action.

Escape-resistance assessment

Assess evidence for resisting escape from a defined execution environment. The review connects runtime versions, privilege configuration, host interfaces and an authorised threat model to a bounded review of execution boundaries, giving the runtime owner a clearer basis for the next action.

Filesystem-access testing

Verify what files an agent execution environment can read or change. The review connects mount maps, path permissions, symlink rules and synthetic sensitive-file markers to a bounded review of execution boundaries, giving the runtime owner a clearer basis for the next action.