diff --git a/website/src/content/docs/learning-hub/building-custom-agents.md b/website/src/content/docs/learning-hub/building-custom-agents.md index ffbdc7ef3..91157e61c 100644 --- a/website/src/content/docs/learning-hub/building-custom-agents.md +++ b/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: 2026-08-09 +lastUpdated: 2026-09-04 estimatedReadingTime: '10 minutes' tags: - agents @@ -73,6 +73,20 @@ tools: ['codebase', 'terminal', 'github'] **model** (recommended): The AI model that powers the agent. Choose based on the complexity of the task—use more capable models for nuanced reasoning. +**Model fallback lists** *(v1.0.83+)*: `model` can also be an array of models, tried in order until one is available to you—useful when an agent's preferred model is temporarily rate-limited or not enabled for your account: + +```yaml +--- +name: 'Security Reviewer' +description: 'Expert security auditor for OWASP vulnerabilities' +model: ['Claude Sonnet 5', 'Claude Sonnet 4', 'GPT-5.6'] +model-policy: required +tools: ['codebase', 'terminal', 'github'] +--- +``` + +Add `model-policy: required` to keep any manual model changes restricted to that list—so a user can't switch the agent to a model outside the approved set mid-session. + **reasoningEffort** *(v1.0.66+)*: Override the reasoning effort level for this agent. Accepted values are `low`, `medium`, and `high`. This lets you pin specific agents to a cost/quality tradeoff regardless of the user's global setting — for example, a quick code-formatting agent can use `low` effort, while a security reviewer uses `high`: ```yaml diff --git a/website/src/content/docs/learning-hub/copilot-configuration-basics.md b/website/src/content/docs/learning-hub/copilot-configuration-basics.md index f661b3f36..505742c0c 100644 --- a/website/src/content/docs/learning-hub/copilot-configuration-basics.md +++ b/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: 2026-09-01 +lastUpdated: 2026-09-04 estimatedReadingTime: '10 minutes' tags: - configuration @@ -457,7 +457,7 @@ 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. -**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, and **Gemini 3.7 Flash** (v1.0.81+). **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. +**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+), **kimi-k3** (v1.0.79+), and **claude-fable-5.1** (v1.0.83+). **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. Note that older, retired Claude and Gemini models are periodically removed from the `/model` picker as newer versions supersede them (v1.0.83+). **Plan mode model** *(v1.0.74+)*: When using plan mode (which blocks file mutations and keeps changes in a planning phase), you can assign a *separate* model specifically for planning — different from your regular session model. This lets you use a fast, cost-effective model for plan drafting while keeping a more capable model on standby for the implementation phase: diff --git a/website/src/content/docs/learning-hub/understanding-mcp-servers.md b/website/src/content/docs/learning-hub/understanding-mcp-servers.md index bf6165429..9282f0091 100644 --- a/website/src/content/docs/learning-hub/understanding-mcp-servers.md +++ b/website/src/content/docs/learning-hub/understanding-mcp-servers.md @@ -3,7 +3,7 @@ title: 'Understanding MCP Servers' description: 'Learn how Model Context Protocol servers extend GitHub Copilot with access to external tools, databases, and APIs.' authors: - GitHub Copilot Learning Hub Team -lastUpdated: 2026-08-28 +lastUpdated: 2026-09-04 estimatedReadingTime: '8 minutes' tags: - mcp @@ -206,6 +206,7 @@ Pagination support means these RPCs work reliably even when a server exposes a l Some MCP servers require authentication to connect to protected resources. GitHub Copilot CLI supports several authentication approaches: - **OAuth**: MCP servers can use the OAuth flow to authenticate with external services. The CLI handles the browser redirect and token storage automatically. This also works when running in ACP (Agent Coordination Protocol) mode. +- **Client ID Metadata Document (CIMD)** *(v1.0.83+)*: MCP OAuth sign-in now supports CIMD, letting an MCP server resolve OAuth client metadata from a hosted document URL instead of requiring manual client registration—useful for enterprise identity providers that publish client metadata centrally. - **`client_credentials` grant type**: For fully headless environments where no browser is available and no user interaction is possible (such as server-to-server automation or CI pipelines), MCP servers can authenticate using the OAuth `client_credentials` grant type. This enables machine-to-machine authentication without any browser redirect or device code prompt. - **Device code flow (RFC 8628)**: When the CLI runs in a **headless or CI environment** where a browser redirect is not possible, it automatically falls back to the device code flow. You'll see a URL and a code to enter on another device to complete authentication. - **`/mcp auth`**: If a token expires or you need to switch accounts, run `/mcp auth` inside a session. This opens the re-authentication UI for any OAuth-enabled MCP server and supports account switching. You can re-authenticate without restarting the session.