Update Learning Hub with recent Copilot CLI, app, and VS Code features

- Document include-custom-instructions frontmatter field for custom agents (CLI v1.0.86+)
- Document Auto routing tier org policy defaults (CLI v1.0.87+)
- Document worktreePathTemplate setting (CLI v1.0.87+)
- Document local sandbox for Copilot app agent sessions (app v1.1.23+)
- Document running agent sessions in local Dev Containers (VS Code 1.138+)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

website/src/content/docs/learning-hub/agents-and-subagents.md

@@ -3,7 +3,7 @@ title: 'Agents and Subagents'
 description: 'Learn how delegated subagents differ from primary agents, when to use them, and how to launch them in VS Code and Copilot CLI.'
 authors:
   - GitHub Copilot Learning Hub Team
-lastUpdated: 2026-09-12
+lastUpdated: 2026-09-21
 estimatedReadingTime: '9 minutes'
 tags:
   - agents
@@ -202,6 +202,8 @@ This pairs with **improved workspace resolution**: agents can resolve a workspac
 
 **Shareable session links** *(VS Code 1.138+)*: Agent Host sessions and chats now provide browser-addressable links, so you can reopen or share a specific Agents window session directly from another app, a GitHub issue, or terminal output — useful when handing off a delegated subagent's work to a teammate for review. You can also start a new session beside the current one in one step, keeping the current work visible while a second session starts.
 
+**Run agent sessions in local Dev Containers** *(VS Code 1.138+)*: Enable `chat.agentHost.devContainer.enabled` to run an agent session inside a project's local Dev Container instead of directly on your machine. When enabled, folders with a supported Dev Container configuration show a **Use Dev Container** action, so the agent works with exactly the toolchain and dependencies your project defines rather than whatever happens to be installed locally. Docker must be installed and running for this to work.
+
 ## Common questions
 
 **Do users always invoke subagents directly?**

website/src/content/docs/learning-hub/building-custom-agents.md

@@ -3,7 +3,7 @@ title: 'Building Custom Agents'
 description: 'Learn how to create specialized GitHub Copilot agents with custom personas, tool integrations, and domain expertise.'
 authors:
   - GitHub Copilot Learning Hub Team
-lastUpdated: 2026-09-05
+lastUpdated: 2026-09-21
 estimatedReadingTime: '10 minutes'
 tags:
   - agents
@@ -108,6 +108,20 @@ tools: ['codebase', 'terminal', 'github']
 
 For MCP server tools, reference them by server name (e.g., `postgres`, `docker`). See [Understanding MCP Servers](../understanding-mcp-servers/) for details.
 
+**include-custom-instructions** *(v1.0.86+)*: By default, custom agents run with their own persona and don't automatically pick up repository instruction files like `AGENTS.md`, `.github/copilot-instructions.md`, or `CLAUDE.md`. Set `include-custom-instructions: true` in the frontmatter if you want the agent to also honor your repository's standing coding conventions on top of its own persona:
+
+```yaml
+---
+name: 'Security Reviewer'
+description: 'Expert security auditor that reviews code for OWASP vulnerabilities'
+model: Claude Sonnet 4
+tools: ['codebase', 'terminal', 'github']
+include-custom-instructions: true
+---
+```
+
+This is useful when an agent's specialized behavior should still respect team-wide standards (e.g., a security reviewer that also follows your repo's style guide) rather than operating purely on its own instructions.
+
 ### Agent Instructions
 
 After the frontmatter, write Markdown instructions that define the agent's behavior. Structure these clearly:

website/src/content/docs/learning-hub/copilot-configuration-basics.md

@@ -3,7 +3,7 @@ title: 'Copilot Configuration Basics'
 description: 'Learn how to configure GitHub Copilot at user, workspace, and repository levels to optimize your AI-assisted development experience.'
 authors:
   - GitHub Copilot Learning Hub Team
-lastUpdated: 2026-09-12
+lastUpdated: 2026-09-21
 estimatedReadingTime: '10 minutes'
 tags:
   - configuration
@@ -457,6 +457,8 @@ The model picker opens in a **full-screen view** with inline reasoning effort ad
 
 **Auto mode and server-side model routing** (v1.0.43+): When you select **Auto** as your model, the CLI uses server-side model routing for real-time model selection. Instead of locking in a single model at session start, Auto mode evaluates each request and routes it to the most appropriate model dynamically. This means straightforward questions can be handled by a faster model while complex reasoning tasks are automatically escalated — without you needing to switch models manually.
 
+**Auto routing tier defaults and org policy** *(v1.0.87+)*: Both individual users and organizations can now set a startup default for the Auto routing tier — the quality/cost tier Auto mode routes requests through. Organizations can enforce a **strict** policy (locking every session in the org to a specific tier) or a **user-overridable** policy (setting a sensible default that individual users can still change for their own sessions). This gives teams a way to standardize on cost-conscious routing org-wide while leaving room for power users to opt into higher tiers when they need to.
+
 **Model family aliases** (v1.0.64+): Instead of typing a full model name, you can use short family aliases in the model setting: `opus`, `sonnet`, `haiku` (Anthropic), and `gpt`, `gemini` (Google/OpenAI). The CLI resolves the alias to the latest available model in that family. This is especially useful in scripts or configuration files where you want to track the best model in a family without hardcoding a version string. Recent models available include **Claude Opus 5** (v1.0.75+), the latest in Anthropic's Opus family for the most demanding tasks, **Grok 4.5** (v1.0.76+) from xAI, **Gemini 3.7 Flash** (v1.0.81+), **Claude Fable 5.1** (v1.0.83+), and **GPT-6 Astra** (v1.0.84+). **Grok 4.6** (v1.0.81+) also gains support for the `xhigh` reasoning effort level, one step above `high`, for the most demanding reasoning tasks. The `/model picker` also periodically retires older models no longer worth recommending — a recent cleanup removed several deprecated Claude and Gemini entries (v1.0.83+), so don't be surprised if a model you previously pinned disappears from the list.
 
 **Model fallback lists** *(v1.0.83+)*: Custom agents can set `model` to a list of several models instead of a single name. Copilot tries each one in order until it finds one available to your account — useful when your preferred model is temporarily rate-limited or not enrolled. Pair this with `model-policy: required` to keep the agent restricted to that list even if you try to switch models mid-session. See [Building Custom Agents](../building-custom-agents/) for the frontmatter syntax.
@@ -845,6 +847,8 @@ These flags apply only to the current invocation — your persisted sandbox pref
 
 **`worktreeBaseRef` setting** *(v1.0.79-8+)*: Controls whether `/worktree`, `/worktree new`, and the `--worktree` startup flag create the new worktree from `HEAD` or from the remote default branch. All three now default to `HEAD`; previously `--worktree` defaulted to starting from the remote default branch. Set this in `/settings` if you want worktrees to branch from the remote default instead.
 
+**`worktreePathTemplate` setting** *(v1.0.87+)*: Controls where `/worktree`, `/move`, `/new`, and `--worktree` create new worktrees. Set a template such as `~/src/worktrees/{repo}/{branch}` using the placeholders `{repoPath}`, `{repo}`, `{branch}`, and `{branchSlug}`. Leave it unset to keep the current default layout — a `<repo>.worktrees/` folder alongside the repository, with slashes in the branch name flattened to dashes.
+
 > **Breaking change — sandbox network isolation (v1.0.83+)**: On macOS and Linux, sandboxed commands can no longer reach services running on your own machine, including a server the sandboxed command itself starts on `127.0.0.1`. This means test suites that bind a local port will fail inside the sandbox. Turn on **Allow local network** in `/sandbox` to restore access to localhost. On Linux, sandboxing also now requires `slirp4netns`, `nsenter`, `iptables`, `ip6tables`, `iptables-restore`, and `ip6tables-restore` on `PATH` — install these if sandboxed commands start failing to launch. Additionally, Linux sandboxes now restrict network egress to the configured HTTP(S) proxy when one is set; this proxy mode requires `slirp4netns`, `util-linux` 2.35+, `iptables`, and `/dev/net/tun` access.
 
 The `--attachment` flag (available in prompt mode, `-p`) lets you attach files — images or native documents — to the initial prompt in non-interactive mode:

website/src/content/docs/learning-hub/github-copilot-app.md

@@ -3,7 +3,7 @@ title: 'Getting Started with the GitHub Copilot app'
 description: 'Learn about the GitHub Copilot app, a desktop experience built for agent-native development. Understand its key features and who it''s for.'
 authors:
   - GitHub Copilot Learning Hub Team
-lastUpdated: 2026-09-12
+lastUpdated: 2026-09-21
 estimatedReadingTime: '8 minutes'
 tags:
   - copilot-app
@@ -127,6 +127,10 @@ Open app settings directly from the message composer with **`/settings`** *(v1.1
 
 > **Terminology note (v1.1.20+)**: The "Start from scratch" option in session creation menus and project pickers has been renamed to **"Chat"**.
 
+### Local Sandbox for Agent Sessions
+
+*(v1.1.23+)* A new project setting and **`/sandbox`** command let you run a local session's shell commands inside a sandbox restricted to that session's workspace, limiting the agent's filesystem access to the project directory it's working in. This gives local sessions the same kind of workspace-scoped isolation that cloud and remote sessions already have by default, without needing a container or VM—useful when you want an extra layer of safety for agents running commands on your own machine.
+
 ## Who is the Copilot app for?
 
 The Copilot app isn't a replacement for existing Copilot experiences—it's another tool in the toolbox. Here's who it serves best: