NVIDIA · miyoungc · Jul 18, 2026 · Jul 17, 2026 · Jul 17, 2026 · Jul 18, 2026
diff --git a/docs/AGENTS.md b/docs/AGENTS.md
@@ -36,6 +36,7 @@ Treat `docs/` as the source of truth for published content and AI-agent Markdown
 - Use `$$nemoclaw` for host CLI command examples on shared OpenClaw, Hermes, and Deep Agents pages.
 - Use literal command names on pages that have only one agent variant.
 - Use `<AgentOnly>` blocks only when content differs by behavior, setup flow, state layout, or agent-specific wording.
+- Treat `<AgentOnly>` as a non-nested build-time directive with opening and closing tags at the first column on their own lines; do not import a runtime component for it.
 - Use route-style links without `.mdx` extensions for links between docs pages.
 - Update `docs/index.yml` when navigation, slugs, or page placement changes.
 

diff --git a/docs/CONTRIBUTING.md b/docs/CONTRIBUTING.md
@@ -178,6 +178,9 @@ Use literal command names on those single-variant pages rather than `$$nemoclaw`
 Run `npm run docs:sync-agent-variants` after editing shared variant source pages or navigation.
 Run `npm run docs` before opening a PR to verify the generated pages, rewritten relative links, and Fern navigation.
 If content differs by behavior, setup flow, state layout, or agent-specific wording, keep using `<AgentOnly>` blocks for that content.
+Treat `<AgentOnly>` as a build-time directive rather than a React component, and do not import it from `AgentGuide.tsx`.
+Put each opening and closing tag at the first column on its own line, and do not nest the blocks.
+The generated pages must contain only statically resolved content, with no `AgentGuide` imports or runtime agent components.
 
 ## Route-Style Links
 

diff --git a/docs/_components/AgentGuide.tsx b/docs/_components/AgentGuide.tsx
diff --git a/docs/about/how-it-works.mdx b/docs/about/how-it-works.mdx
@@ -9,8 +9,6 @@ keywords: ["how nemoclaw works", "nemoclaw sandbox lifecycle blueprint"]
 content:
   type: "concept"
 ---
-import { AgentCli, AgentOnly } from "../_components/AgentGuide";
-
 This page explains how NemoClaw runs supported agents inside an OpenShell sandbox and how the gateway connects the agent to inference, integrations, and policy.
 
 NemoClaw does not replace OpenShell or your chosen agent runtime.
@@ -106,16 +104,21 @@ NemoClaw follows these architecture principles.
 
 Versioned blueprint
 : Host-side orchestration uses a versioned blueprint and runner that can evolve on its own release cadence.
-<AgentOnly variant="openclaw"> The OpenClaw sandbox plugin stays small and stable inside the container.</AgentOnly>
+
+<AgentOnly variant="openclaw">
+
+The OpenClaw sandbox plugin stays small and stable inside the container.
+
+</AgentOnly>
 
 Respect CLI boundaries
-: The <AgentCli /> CLI is the primary interface for sandbox management.
+: The `$$nemoclaw` CLI is the primary interface for sandbox management.
 
 Supply chain safety
 : Blueprint artifacts are immutable, versioned, and digest-verified before execution.
 
 OpenShell-backed lifecycle
-: NemoClaw orchestrates OpenShell resources under the hood, but <AgentCli /> onboard is the supported operator entry point for creating or recreating NemoClaw-managed sandboxes.
+: NemoClaw orchestrates OpenShell resources under the hood, but `$$nemoclaw onboard` is the supported operator entry point for creating or recreating NemoClaw-managed sandboxes.
 
 Reproducible setup
 : Running setup again recreates the sandbox from the same blueprint and policy definitions.
@@ -149,7 +152,7 @@ This separation keeps agent-specific sandbox assets focused and lets host orches
 
 ## Sandbox Creation
 
-When you run <AgentCli /> onboard, NemoClaw creates an OpenShell sandbox that runs your selected agent in an isolated container.
+When you run `$$nemoclaw onboard`, NemoClaw creates an OpenShell sandbox that runs your selected agent in an isolated container.
 The host CLI and blueprint runner orchestrate this process through the OpenShell CLI:
 
 1. NemoClaw resolves the blueprint, checks version compatibility, and verifies the digest.
@@ -166,7 +169,7 @@ During onboarding, NemoClaw validates the selected provider and model, configure
 The sandbox then talks to `inference.local`, while the host owns the actual provider credential and upstream endpoint.
 When you select the Model Router provider, `inference.local` routes to a host-side router that chooses from the configured NVIDIA model pool for each request.
 <AgentOnly variant="hermes">
-For Hermes, <AgentCli /> `inference set` updates `/sandbox/.hermes/config.yaml` at runtime without rebuilding the sandbox.
+For Hermes, `$$nemoclaw inference set` updates `/sandbox/.hermes/config.yaml` at runtime without rebuilding the sandbox.
 </AgentOnly>
 <AgentOnly variant="deepagents">
 For Deep Agents, the managed `dcode` runtime reads the OpenAI-compatible route that NemoClaw writes into `/sandbox/.deepagents/config.toml`.

diff --git a/docs/about/overview.mdx b/docs/about/overview.mdx
@@ -11,8 +11,6 @@ content:
 skill:
   priority: 10
 ---
-import { AgentCli, AgentOnly } from "../_components/AgentGuide";
-
 <AgentOnly variant="openclaw,hermes">
 
 NVIDIA NemoClaw is an open-source reference stack for running always-on AI agents more safely inside OpenShell containers.
@@ -31,7 +29,7 @@ These controls help agents run in clouds, on-premises environments, RTX PCs, and
 NemoClaw pairs hosted inference providers or local model endpoints with a hardened sandbox, routed inference, and declarative egress policy.
 This keeps deployments repeatable and easier to constrain.
 The sandbox runtime comes from [NVIDIA OpenShell](https://github.com/NVIDIA/OpenShell).
-NemoClaw adds the blueprint, <AgentCli /> CLI, onboarding, and related tooling as the reference way to run supported agents there.
+NemoClaw adds the blueprint, `$$nemoclaw` CLI, onboarding, and related tooling as the reference way to run supported agents there.
 
 | Capability              | Description                                                                                                                                          |
 |-------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------|
@@ -72,7 +70,7 @@ NemoClaw provides these benefits to mitigate those risks.
 | Sandboxed execution        | Every agent runs inside an OpenShell sandbox with Landlock, seccomp, and network namespace isolation. The sandbox grants no access by default. |
 | Routed inference           | The OpenShell gateway routes model traffic to your selected provider, transparent to the agent. You can switch providers or models. Refer to [Choose an Inference Provider](../inference/learn-and-choose/choose-inference-provider).          |
 | Declarative network policy | YAML defines egress rules. OpenShell blocks unknown hosts and surfaces them to the operator for approval.                 |
-| Single CLI                 | The <AgentCli /> command orchestrates the full stack: gateway, sandbox, inference provider, and network policy.           |
+| Single CLI                 | The `$$nemoclaw` command orchestrates the full stack: gateway, sandbox, inference provider, and network policy.           |
 | Blueprint lifecycle        | Versioned blueprints handle sandbox creation, digest verification, and reproducible setup.                             |
 
 ## Use Cases

diff --git a/docs/configure-agents/progressive-tool-disclosure.mdx b/docs/configure-agents/progressive-tool-disclosure.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw tool disclosure", "progressive tool disclosure", "search_to
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Progressive tool disclosure limits the tool schemas placed in the model's initial context.
 Each supported agent keeps its native discovery mechanism and configuration schema.
 

diff --git a/docs/deployment/set-up-mcp-bridge.mdx b/docs/deployment/set-up-mcp-bridge.mdx
@@ -11,8 +11,6 @@ content:
 skill:
   priority: 30
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 NemoClaw lets a sandboxed agent use authenticated Streamable HTTP MCP servers without copying external service credentials into the sandbox.
 
 The integration has three parts:

diff --git a/docs/get-started/prerequisites.mdx b/docs/get-started/prerequisites.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw prerequisites", "nemoclaw supported platforms", "nemoclaw h
 content:
   type: "reference"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Before you start, verify that your machine has the software and hardware needed to run NemoClaw.
 
 ## Hardware

diff --git a/docs/get-started/windows-preparation.mdx b/docs/get-started/windows-preparation.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw windows wsl2 setup", "nemoclaw install windows docker deskt
 content:
   type: "reference"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Run NemoClaw inside Windows Subsystem for Linux (WSL 2) on Windows.
 <AgentOnly variant="openclaw">
 Complete these steps before following the [Quickstart](../quickstart).

diff --git a/docs/inference/choose-compatible-inference-api.mdx b/docs/inference/choose-compatible-inference-api.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw preferred api", "openai responses api", "chat completions e
 content:
   type: "concept"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Custom OpenAI-compatible endpoints use `/v1/chat/completions` at runtime by default.
 Choose the Responses API only when your endpoint implements the required streaming and tool-calling behavior.
 

diff --git a/docs/inference/choose-inference-provider.mdx b/docs/inference/choose-inference-provider.mdx
@@ -9,8 +9,6 @@ keywords: ["choose inference provider", "nemoclaw providers", "hosted local infe
 content:
   type: "concept"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Choose a provider based on where the model runs, which API it exposes, and which credential you can supply.
 NemoClaw validates the provider and model before it creates the sandbox.
 

diff --git a/docs/inference/choose-local-inference-server.mdx b/docs/inference/choose-local-inference-server.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw local inference", "ollama vllm nim", "local inference serve
 content:
   type: "concept"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 NemoClaw supports Ollama, vLLM, and NVIDIA NIM as local inference servers.
 Choose the option that matches your host, model, and operational needs.
 

diff --git a/docs/inference/configure-inference-timeouts.mdx b/docs/inference/configure-inference-timeouts.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw inference timeout", "local inference timeout", "sandbox rea
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 NemoClaw uses separate time budgets for agent requests, local provider validation, and sandbox readiness.
 Change the budget that matches the phase that times out.
 

diff --git a/docs/inference/configure-model-limits.mdx b/docs/inference/configure-model-limits.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw context window", "nemoclaw max tokens", "model limits"]
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Configure model limits before onboarding so NemoClaw can bake them into the sandbox image.
 Changing a build-time model limit on an existing sandbox requires fresh recreation.
 

diff --git a/docs/inference/how-inference-routing-works.mdx b/docs/inference/how-inference-routing-works.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw inference routing", "inference.local", "openshell inference
 content:
   type: "concept"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 NemoClaw gives agents one managed inference route while OpenShell handles the selected upstream provider on the host.
 This design keeps provider selection and credentials outside the sandbox.
 

diff --git a/docs/inference/model-capability-audit.mdx b/docs/inference/model-capability-audit.mdx
@@ -15,7 +15,6 @@ audience: ["maintainers", "contributors"]
 status: "maintained"
 exclude-from-skills-gen: true
 ---
-import { AgentOnly } from "../_components/AgentGuide";
 Use this matrix to maintain model and provider audit evidence for NemoClaw agent behavior.
 Use it to determine whether a supported model works as an agent model, not only whether it can answer a one-shot chat prompt.
 

diff --git a/docs/inference/set-up-anthropic-compatible-endpoint.mdx b/docs/inference/set-up-anthropic-compatible-endpoint.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw anthropic compatible", "custom anthropic endpoint", "anthro
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Use the custom Anthropic-compatible provider to configure a custom base URL and model with `COMPATIBLE_ANTHROPIC_API_KEY`.
 The runtime API differs by agent capability.
 

diff --git a/docs/inference/set-up-ollama.mdx b/docs/inference/set-up-ollama.mdx
@@ -11,8 +11,6 @@ content:
 skill:
   priority: 10
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Use Ollama when you want the default local inference setup path.
 NemoClaw detects Ollama on the host and can install, start, or upgrade it on supported systems.
 

diff --git a/docs/inference/switch-models.mdx b/docs/inference/switch-models.mdx
@@ -9,8 +9,6 @@ keywords: ["switch nemoclaw model", "change inference model", "nemoclaw inferenc
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Change the model on an existing provider route with a runtime update or a fresh sandbox recreation.
 The command requires both the provider and model even when the provider does not change.
 

diff --git a/docs/inference/switch-providers.mdx b/docs/inference/switch-providers.mdx
@@ -9,8 +9,6 @@ keywords: ["switch nemoclaw provider", "change inference provider", "nemoclaw in
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Move a sandbox to another provider family while keeping the OpenShell route, agent configuration, and host registry aligned.
 Use onboarding first when the target provider is not registered.
 

diff --git a/docs/inference/understand-provider-validation.mdx b/docs/inference/understand-provider-validation.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw provider validation", "inference validation", "compatible e
 content:
   type: "concept"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 NemoClaw validates the selected provider and model before it creates a sandbox.
 The exact request depends on the provider API that the agent uses.
 

diff --git a/docs/inference/verify-inference-route.mdx b/docs/inference/verify-inference-route.mdx
@@ -9,8 +9,6 @@ keywords: ["verify inference.local", "nemoclaw inference route", "sandbox infere
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Verify inference through the same `inference.local` path that the agent uses inside the sandbox.
 Reading the active route confirms configuration, but it does not authenticate a model request.
 

diff --git a/docs/manage-sandboxes/add-channels-after-onboarding.mdx b/docs/manage-sandboxes/add-channels-after-onboarding.mdx
@@ -9,8 +9,6 @@ keywords: ["nemoclaw channels add", "add messaging channel", "channels list"]
 content:
   type: "how_to"
 ---
-import { AgentOnly } from "../_components/AgentGuide";
-
 Run channel commands from the host, not from inside the sandbox.
 
 ## Select and Add a Channel