[Agents, Containers] Organize sandbox and workspace docs

public/__redirects

@@ -240,6 +240,26 @@
 /aegis/analytics/ /smart-shield/configuration/dedicated-egress-ips/ips-utilization/ 301
 
 # agents
+/agents/tools/code-execution/ /agents/sandboxes-and-workspaces/ 301
+/agents/tools/code-execution/sandbox/ /agents/sandboxes-and-workspaces/sandbox/ 301
+/agents/tools/persistent-files/ /agents/sandboxes-and-workspaces/persistent-files/ 301
+/agents/tools/code-execution/run-javascript/ /agents/sandboxes-and-workspaces/run-javascript/ 301
+/agents/tools/code-execution/persistent-files/ /agents/sandboxes-and-workspaces/computer-container-backend/ 301
+/agents/tools/persistent-files/persistence/ /agents/sandboxes-and-workspaces/persistence/ 301
+/agents/tools/persistent-files/reference/ /agents/sandboxes-and-workspaces/reference/ 301
+/agents/tools/codemode/ai-sdk/ /agents/tools/codemode/integrations/ai-sdk/ 301
+/agents/tools/codemode/tanstack-ai/ /agents/tools/codemode/integrations/tanstack-ai/ 301
+/agents/tools/codemode/browser/ /agents/tools/codemode/integrations/browser/ 301
+/agents/tools/codemode/mcp/ /agents/tools/codemode/integrations/mcp/ 301
+/agents/tools/codemode/openapi/ /agents/tools/codemode/integrations/openapi/ 301
+/agents/tools/computer-container/ /agents/sandboxes-and-workspaces/computer-container-backend/ 301
+/agents/tools/files-and-code/ /agents/sandboxes-and-workspaces/ 301
+/agents/tools/files-and-code/persistent-files/ /agents/sandboxes-and-workspaces/persistent-files/ 301
+/agents/tools/files-and-code/persistence/ /agents/sandboxes-and-workspaces/persistence/ 301
+/agents/tools/files-and-code/reference/ /agents/sandboxes-and-workspaces/reference/ 301
+/agents/tools/files-and-code/run-javascript/ /agents/sandboxes-and-workspaces/run-javascript/ 301
+/agents/tools/files-and-code/run-linux/ /agents/sandboxes-and-workspaces/computer-container-backend/ 301
+/agents/tools/sandbox/ /agents/sandboxes-and-workspaces/sandbox/ 301
 /agents/communication-channels/ /agents/agent-class/communication-channels/ 301
 /agents/communication-channels/chat/ /agents/agent-class/communication-channels/chat/ 301
 /agents/communication-channels/chat/autonomous-responses/ /agents/agent-class/communication-channels/chat/autonomous-responses/ 301

src/components/cf/AgentsPlatformDiagram.astro

@@ -37,7 +37,11 @@ const runtime = [
 ];
 
 const tools = [
-	{ type: "Sandbox", name: "Bash/Shell", href: "/agents/tools/sandbox/" },
+	{
+		type: "Sandbox",
+		name: "Bash/Shell",
+		href: "/agents/sandboxes-and-workspaces/",
+	},
 	{ type: "MCP", name: "Servers", href: "/agents/tools/mcp/" },
 	{ type: "Browser", name: "CDP", href: "/agents/tools/browser/" },
 	{ type: "AI Search", name: "Retrieval", href: "/agents/tools/ai-search/" },

src/content/changelog/agents/2026-07-22-mcp-codemode-updates.mdx

@@ -32,7 +32,7 @@ export class MyAgent extends Think<Env> {
 
 This setting skips Think's automatic `getAITools()` call. MCP registration, restoration, discovery, raw catalog access, direct calls, and Code Mode connectors continue to work.
 
-Use [`listTools()`](/agents/model-context-protocol/apis/client-api/#thismcplisttools) when you only need the raw MCP catalog. For connector setup, refer to [Use MCP tools with Code Mode](/agents/tools/codemode/mcp/).
+Use [`listTools()`](/agents/model-context-protocol/apis/client-api/#thismcplisttools) when you only need the raw MCP catalog. For connector setup, refer to [Use MCP tools with Code Mode](/agents/tools/codemode/integrations/mcp/).
 
 ## Invoke the Code Mode runtime without the AI SDK
 

src/content/changelog/agents/2026-10-26-computer-durable-javascript.mdx

@@ -37,15 +37,15 @@ export class MyAgent extends DurableObject<Env> {
 
 </TypeScriptExample>
 
-When recovery is triggered, the JavaScript runs again from the beginning, reusing recorded results for completed calls without printing previously recorded output again. If a call's result was not recorded before the interruption, recovery may repeat the call — even if its action already succeeded. For retry handling, including `unsafeToRepeat()` for operations that cannot safely repeat, refer to [Calls that must not run twice](/agents/tools/filesystem/#calls-that-must-not-run-twice).
+When recovery is triggered, the JavaScript runs again from the beginning, reusing recorded results for completed calls without printing previously recorded output again. If a call's result was not recorded before the interruption, recovery may repeat the call — even if its action already succeeded. For retry handling, including `unsafeToRepeat()` for operations that cannot safely repeat, refer to [Calls that must not run twice](/agents/sandboxes-and-workspaces/run-javascript/#calls-that-must-not-run-twice).
 
 Recovery starts on the next call to `workspace.runtime`. To recover without waiting for traffic, call `workspace.runtime.recover()` from a [Durable Object alarm](/durable-objects/api/alarms/).
 
 ## Wait for approvals
 
 ![A run pauses at an approval gate, then resumes when approved or fails when rejected](~/assets/images/agents/computer-approval-flow.svg)
 
-Expose your application functions through [custom modules](/agents/tools/filesystem/#give-code-access-to-your-own-functions). A function can call `ctx.interrupt()` to pause the run until your application supplies an approval decision.
+Expose your application functions through [custom modules](/agents/sandboxes-and-workspaces/run-javascript/#give-code-access-to-your-own-functions). A function can call `ctx.interrupt()` to pause the run until your application supplies an approval decision.
 
 ### Require approval before posting a comment
 
@@ -103,4 +103,4 @@ for (const call of await workspace.runtime.pending()) {
 
 The run continues once all pending calls are resolved.
 
-For setup, current limitations, and how replay works, refer to [Resume runs after a restart](/agents/tools/filesystem/#resume-runs-after-a-restart).
+For setup, current limitations, and how replay works, refer to [Resume runs after a restart](/agents/sandboxes-and-workspaces/run-javascript/#resume-runs-after-a-restart).

src/content/changelog/sandbox/2026-09-30-sandbox-sdk-1-0.mdx

@@ -20,7 +20,7 @@ With 1.0, your class can:
 - Serve [previews](/sandbox/previews/) from ports in the sandbox, with your own hostnames and authentication.
 - Handle outbound requests for each hostname in Worker code, so [credentials](/sandbox/network/call-an-authenticated-api/) and bindings stay in your Worker.
 - Expose only the methods that you want callers to use.
-- Run an [agent](/agents/tools/sandbox/) in the same Durable Object, and give the model a tool that runs commands in the sandbox.
+- Run an [agent](/agents/sandboxes-and-workspaces/sandbox/) in the same Durable Object, and give the model a tool that runs commands in the sandbox.
 
 <TypeScriptExample filename="src/index.ts">
 

src/content/docs/agents/agent-class/communication-channels/chat/chat-agents.mdx

@@ -979,6 +979,8 @@ export class ChatAgent extends AIChatAgent {
 
 </TypeScriptExample>
 
+To let the model combine these tools by writing JavaScript, use the [Code Mode AI SDK integration](/agents/tools/codemode/integrations/ai-sdk/).
+
 ### Client-side tools
 
 Define a tool on the server without `execute`, then handle it on the client with `onToolCall`. Use this for tools that need browser APIs.

src/content/docs/agents/concepts/tools.mdx

@@ -66,7 +66,7 @@ The [Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction)
 An Agent can expose MCP tools through either model interface:
 
 - Pass MCP tools directly to a model with the [Agents MCP client](/agents/tools/mcp/).
-- Expose MCP tools inside Code Mode for composition and progressive discovery with [MCP connectors](/agents/tools/codemode/mcp/).
+- Expose MCP tools inside Code Mode for composition and progressive discovery with [MCP connectors](/agents/tools/codemode/integrations/mcp/).
 
 An MCP server can also expose Code Mode itself. For example, it can present one `code` tool or separate `search` and `execute` tools. For these server-side patterns, refer to [Build a Code Mode MCP server](/agents/model-context-protocol/guides/build-codemode-mcp-server/).
 

src/content/docs/agents/harnesses/think/tools.mdx

@@ -272,7 +272,7 @@ export class MyAgent extends Think<Env> {
 
 `includeMcpTools` controls only the automatic model tool merge. MCP connections still register, restore, discover, and wait. Raw catalog access, direct calls, Code Mode connectors, and explicit `this.mcp.getAITools()` calls continue to work.
 
-Use this property instead of removing MCP tool names through `activeTools` in `beforeTurn`. Think converts MCP schemas before it calls `beforeTurn`, so `activeTools` cannot avoid that conversion. To configure a connector runtime, refer to [Use MCP tools with Code Mode](/agents/tools/codemode/mcp/).
+Use this property instead of removing MCP tool names through `activeTools` in `beforeTurn`. Think converts MCP schemas before it calls `beforeTurn`, so `activeTools` cannot avoid that conversion. To configure a connector runtime, refer to [Use MCP tools with Code Mode](/agents/tools/codemode/integrations/mcp/).
 
 Add MCP servers programmatically or via `@callable` methods:
 

src/content/docs/agents/index.mdx

@@ -11,7 +11,12 @@ head:
     content: Agents
 ---
 
-import { PackageManagers, RelatedProduct } from "~/components";
+import {
+	CardGrid,
+	LinkCard,
+	PackageManagers,
+	RelatedProduct,
+} from "~/components";
 import AgentsPlatformDiagram from "~/components/cf/AgentsPlatformDiagram.astro";
 
 Build and host Agents on Cloudflare, connect chat, voice, email, Slack, and webhooks to a durable agent runtime with Browser, Sandbox, AI Search, MCP, Payments, and other MCP tools.
@@ -27,7 +32,20 @@ Agents on Cloudflare are composed from four parts:
 - **Communication channels** define how users and systems reach your agent, such as [chat](/agents/agent-class/communication-channels/chat/), [voice](/agents/agent-class/communication-channels/voice/), [email](/agents/agent-class/communication-channels/email/), [Slack](/agents/agent-class/communication-channels/slack/), [webhooks](/agents/agent-class/communication-channels/webhooks/), and other event sources.
 - **The agent harness** defines the loop: how the agent calls models, selects tools, handles tool results, streams responses, and decides whether to continue. Use the [Pi harness](/agents/harnesses/pi/) to run Pi Durable in a Durable Object, or build your own loop directly on the [Agents SDK runtime](/agents/agent-class/).
 - **The Agents SDK runtime** provides durable infrastructure: the [`Agent` class](/agents/agent-class/internals/), [state](/agents/agent-class/state/), [sessions](/agents/runtime/lifecycle/sessions/), [routing](/agents/agent-class/routing/), [WebSockets](/agents/agent-class/websockets/), [scheduling](/agents/agent-class/schedule-tasks/), [fibers](/agents/agent-class/durable-execution/), and [observability](/agents/runtime/operations/observability/).
-- **Tools** give the agent capabilities: [browser automation](/agents/tools/browser/), [sandboxed code execution](/agents/tools/sandbox/), [AI Search](/agents/tools/ai-search/), [MCP tools](/agents/tools/mcp/), and [payments](/agents/tools/payments/). [Code Mode](/agents/tools/codemode/) lets models discover and orchestrate multiple tools by writing code.
+- **Tools** give the agent capabilities: [browser automation](/agents/tools/browser/), [sandboxed code execution](/agents/sandboxes-and-workspaces/sandbox/), [AI Search](/agents/tools/ai-search/), [MCP tools](/agents/tools/mcp/), and [payments](/agents/tools/payments/). [Code Mode](/agents/tools/codemode/) lets models discover and orchestrate multiple tools by writing code.
+
+<CardGrid>
+	<LinkCard
+		title="Sandboxes and workspaces"
+		href="/agents/sandboxes-and-workspaces/"
+		description="Give your agent persistent files and an isolated environment to run code."
+	/>
+	<LinkCard
+		title="Code Mode"
+		href="/agents/tools/codemode/"
+		description="Let your agent combine tool calls by writing code."
+	/>
+</CardGrid>
 
 ### Get started