Network Policies
NemoClaw runs with a deny-by-default network policy. The sandbox can only reach endpoints that are explicitly allowed. OpenShell intercepts any request to an unlisted destination and prompts the operator to approve or deny it in real time through the TUI.
Baseline Policy
Deep Agents sandboxes use an agent-specific baseline policy in agents/langchain-deepagents-code/policy-additions.yaml so the managed dcode runtime can reach inference, GitHub, and package endpoints while keeping the same deny-by-default model.
Deep Agents uses strict Landlock compatibility, so sandbox startup fails when OpenShell cannot enforce the managed filesystem policy.
Filesystem
/dev/pts is the pseudo-terminal (devpts) directory.
It is writable so PTY-based tools (tmux, script, and interactive shells) can allocate a terminal.
Without it, those tools fail with fork failed: Permission denied.
Read-only access to /var/lib/dpkg lets dpkg-query inspect installed package metadata.
The filesystem policy does not grant write access to the package database.
The sandbox process runs as a dedicated sandbox user and group.
For Deep Agents, Landlock enforcement is strict.
If the kernel or runtime cannot enforce the managed filesystem policy, sandbox startup fails closed.
Network Policies
The following endpoint groups are allowed by default:
The separate raw.githubusercontent.com route lets Deep Agents Code follow GitHub file links and read repository source through its managed fetch_url tool.
Repository, ref, and file path segments vary by task, so the route covers the host while limiting requests to read-only GET and HEAD methods and the listed managed binaries.
All endpoints use TLS termination and are enforced at port 443.
Deep Agents is a terminal-oriented harness.
NemoClaw does not configure messaging channel bridges for Deep Agents today, and the baseline policy does not include Tavily, LangSmith, MCP hosts, or arbitrary web endpoints.
Apply maintained presets such as github, pypi, or tavily only when the sandbox needs that access.
Policy Tiers
During onboarding, the wizard prompts for a policy tier that determines the default set of presets applied on top of the baseline policy.
The baseline policy is always applied regardless of the selected tier.
An operator can persistently exclude a specific baseline entry with policy exclude when they accept reduced, unsupported functionality.
This is the supported registry-backed way to replay a live exact-key removal during rebuild; raw openshell policy set edits are not replayed, while edits made to the source baseline itself are applied when the sandbox is recreated.
When Personal is selected or carried forward, the personal-open-internet preset is mandatory for every agent and every onboarding entry point.
Interactive choices, NEMOCLAW_POLICY_MODE=custom, and NEMOCLAW_POLICY_MODE=skip control only additional presets; they cannot deselect, skip, or replace Personal’s required web authority.
Personal Tier Network Access
The Personal tier applies the personal-open-internet policy preset with a hostless L4 endpoint on destination ports 80 and 443.
The policy matches any requested host on either port, then permits the connection only when every resolved address is in the preset’s allowed ranges.
The rule does not inspect the application protocol or payload, so traffic on these ports is not limited to HTTP or HTTPS.
OpenShell does not restrict the hostname, HTTP method, path, or body after the rule permits the connection.
An agent can send workspace data or sandbox-visible credentials to an arbitrary reachable service on either port without an operator approval prompt.
The preset excludes unspecified, loopback, and link-local address ranges, including the common cloud metadata range. OpenShell also keeps its hard blocks for those destinations. Other destination ports remain denied unless another policy entry permits them. The sandbox’s filesystem, process, gateway authentication, and managed credential controls remain active. Use this tier only for trusted personal workloads with trusted prompts and data.
Every fresh onboarding run through the experimental Portable profile selects the Personal tier.
When NEMOCLAW_POLICY_PRESETS is unset, blank, or contains only whitespace, Portable uses suggested mode with no optional preset override.
If NEMOCLAW_POLICY_PRESETS contains a non-blank list, Portable treats that list as authoritative for additional presets while still applying mandatory personal-open-internet.
nemo-deepagents onboard --resume does not override a recorded non-Personal tier; a resumed Personal tier retains or repairs its mandatory preset.
After selecting a tier, a combined preset and access-mode screen lets you include or exclude optional presets and toggle each between read (GET only) and read-write (GET + POST/PUT/PATCH) access.
On Personal, NemoClaw restores personal-open-internet if it is deselected in the screen.
Tier-default presets are pre-selected; additional presets can be added from the built-in preset list available to the sandbox’s active agent.
NemoClaw filters tier defaults and built-in preset choices by the active agent’s supported integrations.
The personal-open-internet preset uses L4 passthrough, so its read-write label does not add HTTP method or path inspection.
Deep Agents can use the maintained tavily opt-in path, but messaging channel presets are omitted because the terminal harness does not run a NemoClaw messaging bridge today.
NemoClaw automatically suggests the preset that matches the selected provider and removes stale web search presets during resume reconciliation when you switch providers or disable web search.
The Personal tier does not select a Brave Search or Tavily Search preset by default.
Its broad route supports ordinary keyless web fetches, while web_search still requires a separately configured provider.
Explicit custom preset lists and manual interactive selections remain operator-controlled for additional presets.
When LangChain Deep Agents Code is onboarded with --observability, NemoClaw adds the observability-otlp-local preset on Balanced, Open, and Personal tiers.
The Restricted tier suppresses this agent-required preset during onboarding and rebuild reconciliation.
An operator can add it manually after reviewing the additional egress, but the next Restricted reconciliation removes it.
The applied set therefore reflects the chosen tier plus any agent-required presets, so policy list may show one or more presets that do not appear in the tier table above.
The policy list provenance tags are inferred from the current tier YAML and the active agent at display time and are not persisted per preset.
A preset whose name matches an entry in the sandbox’s current tier definition is labelled [from <tier> tier] even when an operator added it manually with policy add after onboarding; agent-specific preset names are only labelled [from <agent> agent] when the active agent matches.
Claude Code direct egress is not included in the Restricted, Balanced, or Open tiers.
Personal’s broad route permits its port 443 transport but does not install or configure Claude Code; on other tiers, apply the claude-code preset explicitly if you install and run the CLI inside the sandbox with its own credentials.
Normal NemoClaw Anthropic inference still routes through the OpenShell gateway.
Tier definitions are stored in nemoclaw-blueprint/policies/tiers.yaml.
In non-interactive mode, set the tier with NEMOCLAW_POLICY_TIER:
Unset, blank, or whitespace-only NEMOCLAW_POLICY_TIER values use the balanced default.
In non-interactive onboarding, a non-blank value that does not match a known tier exits before preflight, gateway, or inference side effects and lists the valid options.
Interactive onboarding ignores an invalid environment value and shows the normal tier prompt.
Inference
The baseline policy allows only the local inference route.
External inference providers are reached through the OpenShell gateway, not by direct sandbox egress.
Local OTLP Trace Export
The observability-otlp-local preset supports the opt-in LangChain Deep Agents Code trace path.
It is not a general remote observability policy.
The sandbox sends OTLP/HTTP protobuf traces to a collector that the operator runs on the host. The managed exporter uses standard OTLP transport headers but does not accept operator-supplied custom or authentication headers. Remote backend endpoints and credentials stay in that collector. The policy does not allow direct LangSmith, Jaeger, Phoenix, or other backend egress from Deep Agents Code. Changing the collector’s downstream exporter requires no sandbox policy change.
OpenShell observes Deep Agents Code export as the managed Python interpreter, so this permission is process-wide for /opt/venv/bin/python3* rather than limited to the dcode launcher.
Sandbox Python can forge spans and resource attributes.
The explicit --observability opt-in can export bounded prompts, responses, tool arguments, tool results, and operational metadata.
Managed size and recognized-key redaction do not detect secrets embedded in ordinary content values.
The collector must enforce the operator’s filtering and redaction requirements before forwarding traces, and it must not treat span fields such as service.name as authenticated tenant identity.
For a safe host binding, policy recovery commands, and a runnable collector, refer to Set Up Deep Agents Trace Export.
Operator Approval Flow
When the agent attempts to reach an endpoint not listed in the policy, OpenShell intercepts the request and presents it in the TUI for operator review.
The Personal tier does not prompt for matching TCP connections on destination ports 80 or 443 because personal-open-internet already permits them.
The flow has these steps:
- The agent makes a network request to an unlisted host.
- OpenShell blocks the connection and logs the attempt.
- The TUI command
openshell termdisplays the blocked request with host, port, and requesting binary. - The operator approves or denies the request.
- If approved, the endpoint is added to the running policy for the session.
To monitor requests as the agent runs, open the TUI:
For step-by-step navigation and approval controls, refer to Approve or Deny Agent Network Requests.
Modifying the Policy
Static Changes
Edit agents/langchain-deepagents-code/policy-additions.yaml and re-run the onboard wizard:
Dynamic Changes
Apply policy updates to a running sandbox without restarting:
To replace the live policy with a complete base policy file, export the current base policy and use openshell policy set.
Requires OpenShell 0.0.72+ for the round-trippable policy get --base and policy set --wait syntax.
NemoClaw strips the OpenShell metadata header and exits non-zero if it cannot validate the base policy.
Do not add --raw when you plan to edit and reapply the file.
Edit or review current-policy.yaml, then apply it:
Excluding a baseline entry
The baseline policy applies to every sandbox, but an operator can persistently exclude a specific baseline entry when they accept reduced, unsupported functionality in exchange for tighter egress.
Preview the exact scope, then exclude an exact baseline key:
The command prints every endpoint, method/path rule, and binary the exclusion removes.
It also names the supported features that may stop working before it requires explicit acknowledgement (--force in non-interactive use).
NemoClaw refuses to exclude an entry that does not have a reviewed feature-impact disclosure, so a new baseline entry cannot bypass this review.
The versioned exclusion record is bound to the reviewed entry content and active agent, recorded in the sandbox registry, and replayed on rebuild.
If the sandbox’s active agent changes, NemoClaw requires you to clear or review the exclusion again instead of applying an approval from another agent baseline.
NemoClaw journals the cross-system update before changing the live OpenShell policy and verifies the exact live result before publishing the durable intent.
If the process or persistence layer fails between those steps, policy list, policy explain, and status report repair required.
Sandbox creation or recreation, rebuild, and cross-sandbox snapshot cloning stop before destructive work, and rerunning the same policy exclude or policy restore command reconciles only an exact source or target state.
An unreadable live policy, or one that matches neither the pre-mutation source nor the intended target, remains fail-closed for manual inspection instead of being guessed or overwritten.
An interrupted restore finalizes only when the durable exclusion still exactly matches the staged exclusion and the current release baseline still exactly matches the journaled target; if either changes or becomes unreadable, the journal remains pending for inspection and re-review.
If a later release changes that entry, create and rebuild fail closed until re-review; if the release removes it, they fail closed until the stale record is cleared.
The critical managed_inference entry cannot be excluded because it carries the required route to managed inference.
NemoClaw enforces this both when the command runs and whenever durable exclusion state is replayed.
An excluded baseline key remains reserved while the exclusion is active. NemoClaw refuses onboarding, rebuild, built-in and custom presets, messaging channel policies, and generated MCP policies that would define the same key and silently restore its egress. It also refuses to exclude a key that an applied preset already owns. Restore the baseline entry before applying a preset that intentionally owns the key, or rename a custom entry whose key represents different access.
To recover from baseline drift, first check whether the release changed the entry’s content or removed it entirely (nemo-deepagents <sandbox-name> policy explain or doctor reports which).
If the entry still exists with different content, policy restore <key> --force restores the current baseline entry and clears the stale exclusion record.
This allows the entry’s listed egress before you can exclude it again.
Preview the restore and exclusion scopes before applying either change:
If the release removed the entry entirely, policy exclude <key> fails with Unknown baseline entry '<key>'. because there is nothing left to exclude.
Preview the stale-record cleanup, then clear the record without changing live egress:
List active exclusions with policy list.
policy explain, status, doctor, and snapshot and rebuild summaries also disclose active exclusions and their reduced-support impact.
Status and doctor compare each approval with the active agent baseline and verify that the excluded key is absent from the live OpenShell policy.
An unreadable live policy is unverified, while a live policy that contains the key is a mismatch that requires repair before you rely on the exclusion.
When the current baseline still defines the entry, policy restore <key> --force allows the entry’s listed egress again and clears its exclusion.
When the baseline no longer defines the entry, the command clears only the stale exclusion record.
Preview the applicable result before you apply it:
Both restore paths require acknowledgement unless you use --dry-run.
A run with a terminal on stdin prompts for confirmation when no acknowledgement flag is present.
A run with NEMOCLAW_NON_INTERACTIVE=1, or a run without a terminal on stdin, requires --force, --yes, or -y.
Excluding a baseline entry leaves agent features that depend on it unsupported for that sandbox.