GitHub Copilot changelog
12 items in the render window (newest 100 + all corrections; not this source's all-time total) ·
status healthy · last successful fetch 2026-11-21 19:16 UTC
Feed for this sourceRSS
Gemini 4 Neon now available in GitHub Copilot
Do:Check Copilot settings model policy to confirm whether Gemini 4 Neon should be enabled for your organization.
WHAT CHANGEDGitHub Copilot added Gemini 4 Neon as a selectable model for Pro+, Max, Business, and Enterprise users across all supported IDEs and clients. The model is billed at provider list pricing under usage-based billing. Rollout is gradual.
WHY IT MATTERSTeams using GitHub Copilot on usage-based billing will incur charges at Google's list pricing when this model is selected. Enterprise and Business administrators must review model policy settings, as new models are automatically enabled by default unless the global default is disabled or the model is explicitly blocked.
WHAT TO DOCheck Copilot settings model policy to confirm whether Gemini 4 Neon should be enabled for your organization. If usage-based billing applies, review provider list pricing at the linked Models and pricing page to understand cost implications before the model becomes available to users.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Gemini 4 Neon in GitHub Copilot +https://github.blog/changelog/2026-11-18-gemini-4-neon-in-github-copilot +Wed, 18 Nov 2026 21:49:21 +0000 + +Gemini 4 Neon is now generally available in GitHub Copilot. It is designed for agentic coding, multistep tool use, and knowledge work. According to Google, Gemini 4 Neon delivers quality close to Gemini 4 Carbon at a lower token cost per task. Google also reports that, compared with Gemini 4 Carbon, it starts responding faster and makes fewer unrequested changes on coding tasks. It is also less likely to make unsupported claims about code, tool, or test results. +This model is billed at provider list pricing under usage-based billing. See Models and pricing for GitHub Copilot for details. +Availability in GitHub Copilot +Gemini 4 Neon is available to Copilot Pro+, Max, Business, and Enterprise users. You can select the model in the model picker in: + +Visual Studio Code +Visual Studio +Copilot CLI +GitHub Copilot coding agent +GitHub Copilot app +github.com +GitHub Mobile on iOS and Android +JetBrains IDEs +Xcode +Eclipse + +Rollout will be gradual. Check back soon if you don’t see it yet. +Enabling access +Copilot Enterprise and Copilot Business plan administrators can manage access to Gemini 4 Neon through the model policy in Copilot settings. Under default model enablement, new models are automatically enabled unless an administrator has turned off the global default or explicitly disables this model. +Learn more +To explore all models available in GitHub Copilot, see our documentation on models and get started with Copilot. +Share your feedback +Join the GitHub Community to share your feedback. + +The post Gemini 4 Neon in GitHub Copilot appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New model availability with pricing implications and deployment scope across all major Copilot tiers. |
| signals | new model added to platform · model availability across multiple tiers (Pro+, Max, Business, Enterprise) · pricing change: billed at provider list pricing · model selection capability in multiple IDEs and clients · admin control over model enablement · performance claims vs. predecessor (faster response, fewer unrequested changes) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot adds eight MCP connectors in public preview
Do:If you use Copilot Business or Enterprise, verify the Editor preview features policy is enabled and check whether your MCP allowlist uses wildcards or lists connectors i…
WHAT CHANGEDGitHub added eight MCP server connectors to Copilot: Box, Buildkite, Calendly, ClickUp, Mixpanel, Supabase, Webflow, and Wix. These join more than 25 existing connectors and are available in public preview across all Copilot plans. Copilot Business and Enterprise require admin enablement of the Editor preview features policy. Enterprises with individual MCP allowlists (not wildcards) must add each new connector explicitly.
WHY IT MATTERSTeams using these services can now query and act on them directly from Copilot sessions without switching contexts. Enterprises with restrictive MCP allowlists will block these connectors by default until an admin adds them, preventing developers from using the integrations even after connecting them.
WHAT TO DOIf you use Copilot Business or Enterprise, verify the Editor preview features policy is enabled and check whether your MCP allowlist uses wildcards or lists connectors individually. If individual, add the eight new connector identifiers to the allowlist before developers attempt to connect them.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Connect Supabase, Box, Calendly, and more to Copilot +https://github.blog/changelog/2026-11-18-connect-supabase-box-calendly-and-more-to-copilot +Wed, 18 Nov 2026 22:04:50 +0000 + +You can now connect eight more services to Copilot using GitHub-verified MCP servers, or connectors, available in public preview. They join more than 25 services you can already connect, like Slack, Figma, and Sentry. To see the full list, open Customize in the Copilot app or browse the Copilot customization page. +New services you can connect + +Box: Search and manage content, and use Box AI. +Buildkite: Work with pipelines and builds. +Calendly: Manage scheduling and meetings. +ClickUp: Track tasks and projects. +Mixpanel: Explore product analytics. +Supabase: Work with projects and databases. +Webflow: Manage sites and content. +Wix: Manage sites and business tools. + +For example, you can ask Copilot why your latest Buildkite build failed and have it fix the code in the same session, or have it check your Supabase schema before it writes a migration. +Connect these services once, and you can use them across Copilot. +Connect a service +You connect these services the same way as any other connector: + +Copilot app: Open Customize and select Connect on the service you want. +Copilot CLI: Run /mcp, switch to the Connectors tab, and select Connect on the service you want. +VS Code: Open the Agents window, go to Customizations, and connect the service you want. +github.com: Browse services on the Copilot customization page and select the service you want. + +Availability +Connectors are in public preview for all Copilot plans. With Copilot Business or Copilot Enterprise, an admin needs to enable the Editor preview features policy. If your enterprise’s MCP allowlist lists connectors individually rather than with a wildcard, an admin also needs to add the new connectors to it. Some services also need to be enabled by an admin on their side, such as allowing Box MCP access. +See all supported connectors, and share feedback in our Community discussion. +Happy coding! 💙 + +The post Connect Supabase, Box, Calendly, and more to Copilot appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.85, needed 0.5 |
|---|---|
| triage reasoning | New integrations and connectors expand Copilot's capabilities; teams building on or managing Copilot deployments need to know about available services and admin requirements. |
| signals | eight new connectors added to Copilot platform · integration expansion (Box, Buildkite, Calendly, ClickUp, Mixpanel, Supabase, Webflow, Wix) · public preview availability across all Copilot plans · admin enablement required for Business/Enterprise plans · MCP allowlist configuration requirement for enterprises · future date: 2026-11-18 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot adds team-level plugin defaults for Business and Enterprise
Do:If you manage GitHub Copilot for teams using Business or Enterprise, review your teams' plugin usage and determine which plugins should be team defaults.
WHAT CHANGEDTeam maintainers and organization owners can now set approved Agent Plugins as defaults for organization teams through a new Customize Copilot tab. Teams can browse plugins from built-in marketplaces, enterprise marketplaces, and add marketplaces from GitHub repositories. Selected plugins are enabled by default for direct team members in VS Code, GitHub Copilot app, and Copilot CLI. Members can override team defaults where policy permits, and child teams do not inherit parent team plugin defaults.
WHY IT MATTERSTeams can standardize plugin configurations without requiring each member to perform identical setup. This reduces onboarding friction and ensures consistent tooling across team members while preserving individual override capability within policy boundaries.
WHAT TO DOIf you manage GitHub Copilot for teams using Business or Enterprise, review your teams' plugin usage and determine which plugins should be team defaults. Navigate to your organization team on github.com, select the Customize Copilot tab, and configure approved plugins for each team.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Customize GitHub Copilot for your team +https://github.blog/changelog/2026-11-18-customize-github-copilot-for-your-team +Wed, 18 Nov 2026 16:00:13 +0000 + +You can now share the GitHub Copilot plugins your team relies on without asking every member to repeat the same setup. Team customizations are generally available for organization teams. +Team maintainers and organization owners can select approved Agent Plugins as team defaults. This gives your team a shared starting point between individual customization and organization-wide policy. +Set up plugins for a team +If you’re a team maintainer or organization owner, open your organization team on github.com and select the Customize Copilot tab. From there, you can: + +Browse plugins from the marketplaces built into Copilot, your enterprise’s marketplaces, and any marketplaces the team adds. You can then add individual plugins to the team. +Add a marketplace from a GitHub repository that contains a plugin marketplace. This makes its plugins available to browse. You can add individual plugins to the team to enable them. + +Selected plugins are enabled by default for the team’s direct members in VS Code, the GitHub Copilot app, and Copilot CLI. Members of child teams don’t inherit them. Team defaults are optional. Members can turn a team plugin off in their Copilot client where policy permits. +Your policies still apply +Enterprise and organization policies take precedence over team choices. Copilot clients continue to enforce which marketplaces are allowed and which plugins are required or blocked. Team defaults don’t override those controls. +Team customizations are available for Copilot Business and Copilot Enterprise. +Learn more about plugins for GitHub Copilot and join the discussion in GitHub Community. + +The post Customize GitHub Copilot for your team appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New team-level plugin customization capability with deployment and permission implications for Copilot users. |
| signals | new feature: team-level plugin customization · new capability: team maintainers can set approved Agent Plugins as defaults · new UI: 'Customize Copilot' tab for organization teams · new marketplace integration: teams can add marketplaces from GitHub repositories · scope clarification: feature available for Copilot Business and Copilot Enterprise · permission model: team defaults optional, members can override where policy permits · inheritance rule: child teams do not inherit parent team plugin defaults |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot adds GPT-6.2 Astra and GPT-6.2 Sol models
Do:Check the model picker in your IDE or Copilot interface to confirm availability.
WHAT CHANGEDGitHub Copilot now offers GPT-6.2 Astra and GPT-6.2 Sol, available to Pro+, Max, Business, and Enterprise users across Visual Studio Code, Visual Studio, JetBrains IDEs, Xcode, Eclipse, Copilot CLI, GitHub Copilot coding agent, GitHub Copilot app, github.com, and GitHub Mobile. Both models are billed at provider list pricing under usage-based billing. Rollout is gradual.
WHY IT MATTERSThese models are designed for agentic coding and code review with stronger coding judgment than earlier GPT-6 models. In early testing, GPT-6.2 Astra completed agentic coding tasks using noticeably fewer tokens than GPT-6 Astra, and GPT-6.2 Sol caught nearly twice as many real issues in code review as GPT-6.1 Sol while keeping findings accurate. Token efficiency affects usage-based billing costs.
WHAT TO DOCheck the model picker in your IDE or Copilot interface to confirm availability. If you are a Copilot Enterprise or Business administrator, review the model policy in Copilot settings to verify whether GPT-6.2 models are enabled under your default model enablement configuration.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## OpenAI’s GPT-6.2 Astra and GPT-6.2 Sol now available +https://github.blog/changelog/2026-11-17-openais-gpt-6-2-astra-and-gpt-6-2-sol-now-available +Tue, 17 Nov 2026 20:20:29 +0000 + +OpenAI’s GPT-6.2 Astra and GPT-6.2 Sol are now generally available in GitHub Copilot. Both models are designed for agentic coding and code review, with stronger coding judgment than earlier GPT-6 models. In early testing, both caught significantly more real issues in code review than earlier GPT-6 models while using fewer tokens. + +GPT-6.2 Astra: OpenAI’s latest Astra model for agentic coding and terminal work. In early testing, it caught significantly more issues in code review than GPT-6 Astra and completed agentic coding tasks using noticeably fewer tokens. +GPT-6.2 Sol: OpenAI’s latest Sol model for complex agentic coding and in-depth code review. In early testing, it caught nearly twice as many real issues in code review as GPT-6.1 Sol while keeping its findings accurate. + +These models are billed at provider list pricing under usage-based billing. See Models and pricing for GitHub Copilot for details. +Availability in GitHub Copilot +GPT-6.2 Astra and GPT-6.2 Sol are available to Copilot Pro+, Max, Business, and Enterprise users. +You can select the models in the model picker in: + +Visual Studio Code +Visual Studio +Copilot CLI +GitHub Copilot coding agent +GitHub Copilot app +github.com +GitHub Mobile on iOS and Android +JetBrains IDEs +Xcode +Eclipse + +Rollout will be gradual. Check back soon if you don’t see the models yet. +Enabling access +Copilot Enterprise and Copilot Business plan administrators can manage access to GPT-6.2 models through the model policy in Copilot settings. Under default model enablement, new models are automatically enabled unless an administrator has turned off the global default or explicitly disables these models. +Learn more +To explore all models available in GitHub Copilot, see our documentation on models and get started with Copilot. +Share your feedback +Join the GitHub Community to share your feedback. + +The post OpenAI’s GPT-6.2 Astra and GPT-6.2 Sol now available appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New production models with stated billing and tier availability affect platform capability and cost planning. |
| signals | new models added to platform · models available to specific user tiers (Pro+, Max, Business, Enterprise) · billing model stated (provider list pricing, usage-based) · token efficiency claim (fewer tokens than predecessors) · admin control over model enablement · broad availability across multiple IDEs and clients |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot adds MAI Code 2 and MAI Code 2 Flash models
Do:If you are a Copilot Enterprise or Business administrator, review the model policy in Copilot settings to verify whether MAI Code 2 and MAI Code 2 Flash should be enable…
WHAT CHANGEDGitHub Copilot now includes two new Microsoft AI coding models: MAI Code 2 (available in Pro, Pro+, Max, Business, and Enterprise tiers) and MAI Code 2 Flash (available in Free, Student, Pro, Pro+, Max, Business, and Enterprise tiers). Both models are available to subscribers on annual plans and can be selected in the model picker across all Copilot interfaces including VS Code, Visual Studio, CLI, JetBrains, Xcode, and GitHub Mobile. Both models use usage-based billing.
WHY IT MATTERSMAI Code 2 Flash is now available in the Free tier, expanding model choice for users without paid subscriptions. The usage-based billing model means costs will vary by actual token consumption rather than flat subscription rates. Enterprise and Business administrators must manage access through model policy settings, as new models are enabled by default unless explicitly disabled.
WHAT TO DOIf you are a Copilot Enterprise or Business administrator, review the model policy in Copilot settings to verify whether MAI Code 2 and MAI Code 2 Flash should be enabled for your organization under the default model enablement policy. If you use GitHub Copilot, check the usage-based billing documentation at Models and pricing for GitHub Copilot to understand how these models will affect costs.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## MAI Code 2 and MAI Code 2 Flash in GitHub Copilot +https://github.blog/changelog/2026-11-17-mai-code-2-and-mai-code-2-flash-in-github-copilot +Tue, 17 Nov 2026 18:01:45 +0000 + +The next generation of Microsoft AI (MAI) coding models has arrived in GitHub Copilot. Introducing MAI Code 2 and MAI Code 2 Flash, two new models that represent a major step forward for the MAI-Code family, expanding from lightweight coding assistance to complex, agentic software engineering. + +MAI Code 2: A powerful, general purpose coding model built to help developers get from request to result with focused responses, practical code changes, and less unnecessary back and forth. Ideal for complex tasks like deep debugging, multi-file implementation, refactoring, and repository-wide reasoning. +MAI Code 2 Flash: A lightweight, cost-efficient coding model optimized for everyday development workflows. Well suited for code explanations, targeted edits, routine fixes, and rapid iteration. + +Both models are designed for agentic coding workflows, giving developers more flexibility to choose the model that best fits their task. +Read the MAI Code 2 and MAI Code 2 Flash announcement for a deeper look at the model’s vision support, coding performance improvements, and efficiency gains. +These models are billed under usage-based billing. For more details, see Models and pricing for GitHub Copilot. +Availability in GitHub Copilot + +MAI Code 2: Copilot Pro, Pro+, Max, Business, and Enterprise +MAI Code 2 Flash: Copilot Free, Student, Pro, Pro+, Max, Business, and Enterprise + +Both models are available to eligible GitHub Copilot subscribers on annual plans. +You can select the models in the model picker in: + +Visual Studio Code +Visual Studio +Copilot CLI +GitHub Copilot cloud agent +GitHub Copilot app +github.com +GitHub Mobile iOS and Android +JetBrains +Xcode +Eclipse + +Enabling access +Copilot Enterprise and Copilot Business plan administrators can manage access to MAI models through the model policy in Copilot settings. Under default model enablement, new models are enabled automatically unless an administrator has turned off the global default or explicitly disables the model. +Learn more +To learn more about the new models, see the MAI Code 2 page. +To explore all models available in GitHub Copilot, see our documentation on supported models and get started with Copilot. +Share your feedback +We’d love to hear about your experience using MAI Code 2 and MAI Code 2 Flash. Join the GitHub Community to share your feedback and help us continue improving these models. + +The post MAI Code 2 and MAI Code 2 Flash in GitHub Copilot appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New models added with expanded tier availability and usage-based billing structure affects deployment and cost planning. |
| signals | two new models added to platform · model availability matrix changed (Flash now in Free tier) · pricing model stated: usage-based billing · new agentic workflow capability · admin control over model enablement · future date: 2026-11-17 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot activity now streams to Microsoft Agent 365
Do:If you are a GitHub Enterprise Cloud customer with Enterprise Managed Users and use Microsoft Agent 365, review the streaming audit log documentation on GitHub Docs to d…
WHAT CHANGEDGitHub Enterprise Cloud with Enterprise Managed Users can now stream Copilot agent activity—including prompts, responses, tool calls, and usage across all Copilot clients (GitHub.com, CLI, VS Code, Visual Studio, JetBrains, Eclipse)—directly into Microsoft Agent 365 for centralized audit and review alongside other AI workloads.
WHY IT MATTERSOrganizations using both GitHub Copilot and Microsoft's agent control plane can now consolidate audit data in one location and access risk analysis through Microsoft Purview and Defender, rather than reviewing Copilot activity separately.
WHAT TO DOIf you are a GitHub Enterprise Cloud customer with Enterprise Managed Users and use Microsoft Agent 365, review the streaming audit log documentation on GitHub Docs to determine whether to enable this integration for your compliance or monitoring workflows.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Monitor GitHub Copilot activity through Microsoft Agent 365 +https://github.blog/changelog/2026-11-17-monitor-github-copilot-activity-through-microsoft-agent-365 +Tue, 17 Nov 2026 18:00:59 +0000 + +Enterprises can now stream GitHub Copilot agent activity directly into +Microsoft Agent 365. This lets you bring audit data from across GitHub Copilot clients into Microsoft’s agent control plane, where you can review Copilot activity alongside other AI workloads. +With this release, administrators can: + +See activity across Copilot clients, including cloud agents on GitHub.com, Copilot CLI, Visual Studio Code, Visual Studio, and partner IDEs such as JetBrains and Eclipse. +Review user sessions, including prompts, responses, and tool calls, and view usage trends in Agent 365. +Access additional risk information in Microsoft Purview and Defender to investigate and mitigate risk. + +This is available for GitHub Enterprise Cloud customers with Enterprise Managed Users. To get started, see Streaming the audit log for your enterprise on GitHub Docs. + +The post Monitor GitHub Copilot activity through Microsoft Agent 365 appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New audit streaming integration materially changes monitoring and compliance posture for enterprise customers. |
| signals | new integration capability · audit/monitoring feature addition · cross-client activity visibility · security/compliance feature · enterprise-tier feature · EMU requirement stated |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot Fusion adds three models across tiers
Do:Test Copilot Fusion in your current development environment to confirm the new model routing produces acceptable results for your use cases.
WHAT CHANGEDGitHub added Claude Sonnet 5.5 to Copilot Fusion Max, GPT-6.1 Terra to Copilot Fusion High, and MAI-Code-2 to Copilot Fusion Standard. Copilot Fusion is now available on github.com, Visual Studio, and GitHub Mobile beta apps for iOS and Android, with production mobile app support coming later. Copilot Fusion Max achieved 71.2% on Terminal-Bench 4.1, 5.4 percentage points ahead of Opus 5.5 and Fable 5.2.
WHY IT MATTERSIf you use Copilot Fusion, the model routing logic now includes different models at each tier, which may change response characteristics, latency, or cost for existing workflows. The web and Visual Studio availability means Fusion is no longer limited to specific editor integrations.
WHAT TO DOTest Copilot Fusion in your current development environment to confirm the new model routing produces acceptable results for your use cases. Check whether the model picker now appears in surfaces where you previously could not select Fusion.
Model-reported confidence 75% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## GitHub Copilot Fusion updates: more availability, benchmarks, and models +https://github.blog/changelog/2026-11-17-github-copilot-fusion-updates-more-availability-benchmarks-and-models +Tue, 17 Nov 2026 18:04:48 +0000 + +Today, we’re expanding GitHub Copilot Fusion with new models, available on more Copilot surfaces, and results on two public Terminal-Bench leaderboards, giving you more ways to use Fusion and compare its performance. +GitHub Copilot Fusion joins public Terminal-Bench 4.1 leaderboard +Copilot Fusion is now listed on Terminal-Bench 4.1 public leaderboards hosted by Stanford, Harbor, and the Laude Institute: Terminal-Bench 4.1 measures how well AI agents complete hard, realistic tasks in a terminal. +On Terminal-Bench 4.1, Copilot Fusion Max achieved a 71.2% resolution rate, 5.4 percentage points ahead of leading frontier models Opus 5.5 and Fable 5.2, while costing 88% less than Fable 5.2 and 57% less than Opus 5.5. + +Independent leaderboards let you compare Copilot with other agents on the same tasks, using the same evaluation rules. +Copilot Fusion is now available on web, mobile, and Visual Studio +Copilot Fusion is coming to more places you use Copilot. With this release, you can select Copilot Fusion on the latest versions of github.com, Visual Studio, and GitHub Mobile beta apps for iOS and Android. Support in the production GitHub Mobile apps is coming shortly. +Select Copilot Fusion in the model picker and let it choose and coordinate models for each task. You set the effort level and Copilot handles the rest. +More models added +With this release we are adding: + +Claude Sonnet 5.5 to Copilot Fusion Max +GPT-6.1 Terra to Copilot Fusion High +MAI-Code-2 to Copilot Fusion Standard + +Learn more +Read more about Copilot Fusion. + +The post GitHub Copilot Fusion updates: more availability, benchmarks, and models appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New models added to production service tiers and expanded platform availability affect deployment and integration decisions. |
| signals | new models added to Copilot Fusion tiers · expanded availability across platforms (web, mobile, VS) · public benchmark performance claims · model picker feature availability |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot CLI adds Microsoft Entra authentication in preview
Do:If your organization uses GitHub EMU with OIDC or has Microsoft 365 Copilot licenses, update to GitHub Copilot CLI version 1.0.113 or later by running `copilot update`.
WHAT CHANGEDGitHub Copilot CLI now supports Microsoft Entra authentication in public preview on Windows and macOS (Linux support pending). This enables two distinct modes: (1) silent SSO for GitHub Enterprise Managed Users via OIDC, using GitHub inference and existing Copilot Business/Enterprise licenses, and (2) direct access using Microsoft 365 Copilot licenses with usage-based billing, using Microsoft 365 inference instead of GitHub inference. Minimum version 1.0.113 is required.
WHY IT MATTERSOrganizations using EMU with OIDC will see authentication flow changes—users authenticate via Microsoft Entra instead of GitHub directly. Organizations with Microsoft 365 Copilot licenses and usage-based billing now have an alternative inference path with different billing and model access governed by Microsoft 365 settings rather than GitHub Copilot plans. The inference backend and billing model differ depending on which authentication path is used.
WHAT TO DOIf your organization uses GitHub EMU with OIDC or has Microsoft 365 Copilot licenses, update to GitHub Copilot CLI version 1.0.113 or later by running `copilot update`. Review the GitHub Copilot CLI documentation to determine which authentication mode applies to your setup and verify that the correct inference path and billing model are in effect.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## GitHub Copilot CLI adds Microsoft Entra sign-in as a preview +https://github.blog/changelog/2026-11-17-github-copilot-cli-adds-microsoft-entra-sign-in-as-a-preview +Tue, 17 Nov 2026 18:00:35 +0000 + +GitHub Copilot CLI now supports Microsoft Entra authentication in public preview. You can sign in with your Microsoft Entra account to silently authenticate to a linked GitHub Enterprise Managed User (EMU) account, use your Microsoft 365 Copilot license, or both, depending on your organization’s entitlements. +Microsoft Entra authentication for both EMUs and Microsoft 365 Copilot is currently supported on Windows and macOS. Linux support will follow. +Enterprise Managed User support +For enterprises using EMUs via OIDC, GitHub Copilot CLI now supports silent SSO using Microsoft Entra authentication. You can access GitHub from GitHub Copilot CLI using your Microsoft Entra identity without a separate GitHub sign-in. You must have an assigned Copilot Business or Copilot Enterprise license, and your organization must enable the Copilot CLI policy. In this scenario, GitHub Copilot CLI uses GitHub inference, and usage is governed by the organization’s GitHub Copilot plan. +This SSO experience is currently limited to GitHub Copilot CLI and VS Code and is available only to enterprises using OIDC for EMU. +To enable Enterprise Managed Authorization for your enterprise on GitHub Enterprise Cloud, see the Enterprise Managed Authorization documentation. +Microsoft 365 Copilot license support +If your organization is in Frontier and has a Microsoft 365 Copilot license with usage-based billing, Microsoft Entra authentication lets you access models available through Microsoft Copilot directly from the terminal. In this scenario, GitHub Copilot CLI uses Microsoft 365 inference rather than GitHub inference, and you don’t need a GitHub account. Model access and usage limits are managed through your organization’s Microsoft 365 Copilot settings, not GitHub. +You can also use the Work IQ MCP server to give Copilot access to your Microsoft 365 work data as context. For more information, see the Work IQ MCP overview. +Get started +To take advantage of these capabilities, GitHub Copilot CLI version 1.0.113 or later is required. You can update the CLI by running copilot update in your terminal. +Steps for enabling each capability in the CLI are provided in the GitHub Copilot CLI documentation. +We’d love to hear about your experience using this new Microsoft Entra support. Join the GitHub Community to share your feedback and help us continue improving the experience. + +The post GitHub Copilot CLI adds Microsoft Entra sign-in as a preview appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New authentication method, license type, and inference routing fundamentally alter how teams deploy and bill Copilot CLI. |
| signals | new authentication method added (Microsoft Entra SSO) · new license type supported (Microsoft 365 Copilot) · new inference path introduced (Microsoft 365 inference vs GitHub inference) · platform expansion (Windows, macOS; Linux pending) · breaking change in auth flow for EMU enterprises · usage billing model change (Microsoft 365 usage-based vs GitHub plan) · minimum version requirement (1.0.113+) · new integration capability (Work IQ MCP server) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot app adds Azure DevOps and GHES support, browser use preview
Do:If your team uses Azure DevOps: open Settings > Experimental, enable Azure DevOps integration, then add your Microsoft account under Accounts.
WHAT CHANGEDGitHub Copilot app now supports Azure DevOps repositories and GitHub Enterprise Server accounts, both requiring manual account setup in Settings > Accounts. Browser use entered public preview for Chromium-based browsers (Chrome, Edge) via a new extension, off by default. GHES support moved from opt-in preview to generally available in version 1.2.0 or later. Cloud sessions are unavailable for both Azure DevOps and GHES repositories.
WHY IT MATTERSTeams using Azure DevOps or self-hosted GitHub Enterprise Server can now use Copilot app features (agent branching, PR creation, CI fix automation) that were previously GitHub.com-only. Browser use enables form filling and web app testing in authenticated browser contexts, but requires explicit opt-in and extension installation. The cloud sessions limitation means these repositories cannot use certain Copilot features that depend on cloud execution.
WHAT TO DOIf your team uses Azure DevOps: open Settings > Experimental, enable Azure DevOps integration, then add your Microsoft account under Accounts. If using GHES: update to Copilot app 1.2.0+, then add your GHES account under Settings > Accounts. For browser use: install the GitHub Copilot extension for Chrome/Edge, then enable both Computer Use and Browser Use in Settings > Computer Use.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## GitHub Copilot app, Microsoft Ignite releases +https://github.blog/changelog/2026-11-17-github-copilot-app-microsoft-ignite-releases +Tue, 17 Nov 2026 18:03:42 +0000 + +Copilot is most useful when it can work where you do. The GitHub Copilot app now supports your Azure DevOps repositories, your GitHub Enterprise Server instance, and the browser you’re already signed in to. +Work with Azure DevOps repositories in the GitHub Copilot app +Clone an Azure DevOps repository, ask an agent to create a branch, make code changes, and open a pull request, then see it through to completion without leaving the GitHub Copilot app. Track its status in the sidebar, click Fix on a review comment to have an agent address it, or turn on agent merge to let an agent fix CI failures and complete the pull request once your Azure DevOps checks and policies pass. +Azure DevOps integration is now in public preview. To get started, open Settings > Experimental and turn on Azure DevOps integration, then go to Accounts > Add account and sign in with the Microsoft account you use for Azure DevOps. Cloud sessions aren’t available for Azure DevOps repositories. +For more information, see the Azure DevOps documentation. + +Let Copilot work in your browser +Ask Copilot to fill out forms, pull information from webpages, or test and debug the web app you’re building, right in the browser you’re already signed in to. Browser use, now in public preview, works in Chromium-based browsers, including Google Chrome and Microsoft Edge, with the GitHub Copilot extension for Chrome and runs in an isolated tab group. It also works alongside computer use. +Browser use is off by default. To get started, install the extension, then open Settings > Computer Use and turn on both Enable Computer Use and Enable Browser Use. You can manage permission levels in the extension’s settings. + +Browser use is also available in Copilot CLI, VS Code, and JetBrains IDEs. +For more information, see the browser use documentation. +Bring your GitHub Enterprise Server work into the app +Add your GitHub Enterprise Server (GHES) account to work with remote repositories, issues, and pull requests in the GitHub Copilot app. GHES support, previously an opt-in preview, is now generally available. Update to GitHub Copilot app version 1.2.0 or later to use it. +To get started, sign in with a GitHub.com or GHE.com account, then open Settings > Accounts > Add account and choose GitHub Enterprise Server. Cloud sessions aren’t available for GHES. +For more information, see the Copilot GHES documentation. +Get started +Download or update the GitHub Copilot app. + +The post GitHub Copilot app, Microsoft Ignite releases appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New integrations, feature availability changes, and version requirements affect deployment and capability planning. |
| signals | new integration: Azure DevOps repositories now supported in Copilot app · new feature: browser use in public preview (Chromium-based browsers) · feature promotion: GitHub Enterprise Server support moved from opt-in preview to generally available · version requirement: GHES support requires Copilot app version 1.2.0 or later · deployment change: new settings paths and account configuration workflows · capability constraint: cloud sessions unavailable for Azure DevOps and GHES repositories |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot code review adds adaptive effort selection in public preview
Do:If you use Copilot code review, test Adaptive on a sample of pull requests to verify it selects appropriate effort levels for your codebase's risk profile.
WHAT CHANGEDGitHub Copilot code review now offers Adaptive effort level selection in public preview, which automatically chooses Lite, Balanced, or Max review depth based on each pull request's complexity and risk. Previously, users had to manually select the effort level in settings or when requesting a review. Adaptive can be set as the default in personal, repository, organization, or enterprise settings, and respects configured maximum effort level constraints. Also in public preview: cross-repository context for public repositories, secret detection using both deterministic (500+ known formats) and AI/ML methods, and repository insights tracking. Copilot code review for Azure DevOps moved from public preview to generally available, with usage billed in GitHub AI credits to the linked Azure subscription.
WHY IT MATTERSTeams currently setting review effort levels manually for each pull request or using a single default for all changes can now have the system match review depth to actual risk. This affects review coverage: routine changes get faster feedback, and high-risk changes get deeper analysis without manual triage. For Azure DevOps users, general availability means the feature is production-ready and billing now flows through Azure subscriptions rather than GitHub directly.
WHAT TO DOIf you use Copilot code review, test Adaptive on a sample of pull requests to verify it selects appropriate effort levels for your codebase's risk profile. Check that your configured maximum review effort level allows Adaptive to choose the range you want. For Azure DevOps organizations, confirm your Azure subscription is linked and review GitHub AI credit charges in Azure Cost Management.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Copilot code review: Adaptive effort, secret detection, and more +https://github.blog/changelog/2026-11-17-copilot-code-review-adaptive-effort-secret-detection-and-more +Tue, 17 Nov 2026 18:02:05 +0000 + +GitHub Copilot code review now helps you catch more issues earlier, with less effort. Adaptive matches review depth to each pull request, so routine changes get a quick pass and risky changes get a deeper look. Reviews can also draw on related repositories and flag exposed secrets before they reach your main branch. Features in public preview are subject to change. +Adaptive, cross-repository context, and secret detection are available to Copilot Pro, Pro+, Max, Business, and Enterprise plans. +🎯 Adaptive: The right review effort for each pull request +Today, someone has to decide in settings or when requesting a review, what review effort level each pull request needs. A documentation update may only need Lite, while a change to authentication logic benefits from Max. Adaptive makes that choice for you. Copilot looks at each pull request’s complexity and risk, then reviews it with Lite, Balanced, or Max. Each review shows which effort level Adaptive picked. Adaptive is now in public preview. + +⚙️ Configure Adaptive +You can select Adaptive for a single pull request from the “Reviewers” bar, or set it as your default review effort in personal, repository, organization, or enterprise settings. Balanced remains the default review effort level. + +Adaptive never picks a level above the maximum review effort level set for your enterprise, organization, or repository. To learn more, see Configuring code review by GitHub Copilot. +⚠️ The maximum review effort level configured in Copilot code review settings determines which levels Adaptive can choose from. +🔗 Cross-repository context +Changes often affect code outside the repository you’re working in. With cross-repository context, Copilot code review can read related public repositories during a review, so it can find code in other repositories that depends on your change. It can also find relevant context from documentation stored in other repositories, custom instructions and skills, and more. + +Cross-repository context is now in public preview for public repositories. +🔐 Secret detection in reviews +Copilot code review can now flag exposed secrets in your pull requests. Using both deterministic and non-deterministic approaches, Copilot code review points out exposed credentials in review comments, so you can remove them before you merge. Secret detection in Copilot code review is now in public preview. +Secret detection is supported for the following types: + +Known formats. GitHub collaborates with 150+ technical partners through GitHub’s secret scanning partnership program in order to build out high-confidence secret detectors. These 500+ secret types use deterministic methods like regex to find secrets. +AI/ML-detected secrets. Copilot code review also detects secrets without recognizable format. It uses LLMs and GitHub’s purpose-built model for secret to find exposed secrets. + +To turn on secret detection, go to your repository or organization settings and select Copilot, then Code review. To learn more, see Secret scanning. +📊 Track review outcomes by repository +You can now see how Copilot code review feedback is addressed and utilized in each repository. To get started, navigate to the Insights tab of any repository you have read access to, then click Repository insights. For details, see Track Copilot code review outcomes by repository with Repo Insights. +☁️ Copilot code review for Azure DevOps +Copilot code review for Azure Repos, previously in public preview, is now generally available. Teams that host code in Azure DevOps can get Copilot reviews on their pull requests, with comments and suggested changes on the lines that need attention. + +Flexible setup: Administrators turn on Copilot code review at the organization, project, and repository level. Add a branch policy to request a Copilot review automatically on pull requests that target a branch, or request one manually from the Reviewers list. +Review effort levels: Set a default of Lite or Balanced for a project or repository, or choose a different effort level for a single review. +Custom instructions and skills: Tailor Copilot’s feedback to your team’s standards with custom instructions and agent skills. +Faster follow-up: Apply and commit Copilot’s suggestions and resolve the related comments at the same time. +Admin visibility: Audit log events record when Copilot code review is turned on or off and when its settings change. +Azure billing: Usage is billed in GitHub AI credits to the Azure subscription linked to your Azure DevOps organization, so you can track costs and set budget alerts in Azure Cost Management. + +To get started, see Get started with Copilot code review for pull requests in the Azure DevOps documentation. + +The post Copilot code review: Adaptive effort, secret detection, and more appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | Multiple new capabilities and feature availability changes affecting deployment, security, and billing for engineering teams. |
| signals | new feature: Adaptive review effort (public preview) · new feature: cross-repository context (public preview) · new feature: secret detection in reviews (public preview) · new feature: repository insights tracking · Azure DevOps support now generally available (was public preview) · plan availability stated: Pro, Pro+, Max, Business, Enterprise · billing model for Azure: GitHub AI credits to Azure subscription |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot for JetBrains ends support for IDE 2025.2
Do:Upgrade JetBrains IDE to version 2025.3 or later.
WHAT CHANGEDGitHub Copilot for JetBrains no longer supports JetBrains IDE version 2025.2. Users must upgrade to IDE version 2025.3 or later to continue receiving Copilot updates. GitHub recommends version 2026.2 or later. Additionally, enterprise-managed settings now disable Claude and Codex by default unless an administrator policy allows an exception.
WHY IT MATTERSTeams running JetBrains IDE 2025.2 will stop receiving Copilot updates, including security fixes and compatibility patches. Enterprise users who relied on Claude or Codex models will find them disabled unless their administrator explicitly enables them, potentially breaking existing workflows that depend on those specific models.
WHAT TO DOUpgrade JetBrains IDE to version 2025.3 or later. If you use Claude or Codex models in an enterprise-managed environment, verify with your administrator that these providers are enabled in the policy settings, or adjust workflows to use other available models.
Model-reported confidence 95% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## GitHub Copilot in JetBrains, Ignite releases +https://github.blog/changelog/2026-11-17-github-copilot-in-jetbrains-ignite-releases +Tue, 17 Nov 2026 18:00:59 +0000 + +This update brings remote access, on-device voice dictation, and browser use controls to GitHub Copilot for JetBrains. Copilot users in JetBrains IDEs also get context usage shortcuts, historical diffs, simpler follow-up controls, and improvements to chat reliability. +What’s new +Cloud sandbox (public preview) +Cloud sandbox runs Copilot sessions in an isolated, GitHub-hosted environment instead of on your computer. Copilot can make code changes and run commands and tests there, freeing local resources and allowing parallel sessions. We’re rolling this out in phases. + +Remote access (public preview) +You can now opt in to remote access to make Copilot chats available in other Copilot clients, such as VS Code and the Copilot app. Discover shared sessions from remotely accessible computers, or reopen remote sessions from history to continue the conversation. We’re rolling this out in phases. + +Browser use (public preview) +Paired with GitHub Copilot in Chrome, Copilot’s built-in browser use plugin can open and navigate web pages, read their content, and interact with links, buttons, and forms to gather information or complete web-based tasks. +Enable Browser Use in JetBrains under GitHub Copilot > Chat > Enable Browser Use. + +On-device voice dictation (public preview) +You can now dictate chat messages on supported native plugin packages with Editor Preview enabled. Opt in and download a model to get started. Transcripts stay as editable drafts until you select Send. + +Split mode remote development +GitHub Copilot supports JetBrains split mode remote development, with chat and editor UI running on your local client while Copilot operations run on the remote backend. Make sure you have the GitHub Copilot plugin installed both locally and on the remote backend. +User experience enhancements +You can manage conversations and find model options with fewer steps: + +Conversation timing: See timestamps on Copilot agent prompts and responses. +Session cleanup: Use Delete beside Archive in session history, and choose whether to skip future single-session deletion confirmations. +Model picker: Find Copilot Fusion pinned at the top of Copilot agent model management, where it stays enabled even when you deselect the entire model group. +Organization usage: See Copilot Access usage in the quota dialog consistently with other organization plans, including organization help links rather than upgrade or overage actions. +Steer and queue: Press Enter to steer a running Copilot agent session, or Ctrl+Enter (Command+Enter on macOS) to queue a follow-up. +Inline chat details: See model names and per-response token details while retaining agent session credit totals. Ask mode response footers now show the selected Fusion tier rather than the underlying model, and the model picker stays open when you turn Copilot Fusion on or off. +Worktree names: Use shorter Copilot worktree directory and branch names that retain their timestamps. + +Quality improvements +This update improves chat persistence, startup, and recovery, preserves saved history, and fixes proxy-related loading issues and saved Claude Code paths on Unix. You also get clearer error messages for failed tool calls, working historical session diffs on newer JetBrains IDE EAP builds, and improvements to context search, code formatting, input focus, and MCP sign-in recovery. +Changed +Onboarding and usage tips +The chat onboarding tour is now generally available to eligible users. Chat usage tips are available by default, while your saved tips preference remains unchanged. +Enterprise-managed providers and Copilot Fusion +Enterprise-managed settings now disable Claude and Codex unless an administrator policy allows an exception. Copilot preserves your saved provider preferences. +Copilot Fusion now applies enterprise-managed tier defaults while keeping available tiers selectable and preserving your saved choices. +Code review and Java modernization +Code review now uses the Copilot agent review engine. You still see results in the existing inline review interface. +You can now find Java modernization findings in the “Problems View” on supported IDEs. We’ve restored automatic Maven vulnerability checks: opening or importing a Maven project sends dependency coordinates to GitHub Advisories to check for vulnerabilities. +Deprecation +Support for JetBrains IDE 2025.2 has ended. Upgrade to 2025.3 or later to continue receiving GitHub Copilot updates. We recommend version 2026.2 or later for the best experience. +Try it out +We encourage you to try out the latest version of the GitHub Copilot plugin and share your feedback. We’d love to hear about your experience in the following channels: + +In-product feedback: Use the feedback options within your IDE. +Feedback repository: Share your thoughts in the GitHub Copilot for JetBrains IDEs issues. + +The post GitHub Copilot in JetBrains, Ignite releases appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | IDE version deprecation with upgrade requirement; multiple breaking changes to model availability and execution model. |
| signals | deprecation: JetBrains IDE 2025.2 support ended, upgrade required to 2025.3+ · new feature: cloud sandbox (public preview) changes execution model · new feature: remote access enables cross-client session sharing · new feature: browser use plugin adds new capability · new feature: on-device voice dictation · breaking change: enterprise-managed settings now disable Claude/Codex by default · integration change: code review engine switched to Copilot agent · future date: deprecation effective immediately for 2025.2 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
GitHub Copilot adds strictPluginOnlyCustomization setting for five customization types
Do:Audit current Copilot customizations in user and project configurations across the five categories.
WHAT CHANGEDGitHub Enterprise Cloud administrators can now enable a strictPluginOnlyCustomization managed setting that blocks standalone user and project customizations while allowing plugin-sourced and managed-settings-sourced customizations. The setting applies to five categories: skills, custom agents, hooks, MCP servers, and executable Copilot extensions. It can be applied to all categories or selected individually. Available for Copilot Business and Copilot Enterprise licenses.
WHY IT MATTERSTeams using standalone MCP server configurations, custom hooks, or executable extensions in user or project settings will lose access to those customizations when an administrator enables this restriction. Approved customizations must be redistributed through plugins or managed settings to remain available.
WHAT TO DOAudit current Copilot customizations in user and project configurations across the five categories. Identify which ones are business-critical and coordinate with enterprise administrators to either distribute them through approved plugins or add them to managed settings before the restriction is enabled.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## Control where GitHub Copilot customizations come from +https://github.blog/changelog/2026-11-16-control-where-github-copilot-customizations-come-from +Mon, 16 Nov 2026 19:48:33 +0000 + +As an enterprise administrator, you can now require selected GitHub Copilot customizations to come from plugins or managed settings, rather than standalone user or project configuration. Combined with marketplace restrictions, this lets you distribute approved MCP tools and canvases through a shared plugin library without disabling those capabilities entirely. +Choose which customizations require plugins +The strictPluginOnlyCustomization managed setting blocks standalone customizations from user and project sources. Eligible plugins and managed settings remain allowed, subject to your other policies. +You can apply the restriction to all five supported categories or select individual ones: skills, custom agents, hooks, MCP servers, and executable Copilot extensions. For example, restrict MCP servers and extensions while leaving skills, agents, and hooks open. Extensions here include executable canvas providers, not VS Code Marketplace extensions. +Combine with your marketplace policies +Plugin-only does not mean enterprise-managed plugins only, and it isn’t a marketplace allowlist. Pair it with strictKnownMarketplaces to restrict eligible marketplace sources. Your existing managed settings can supply marketplace definitions and require or block specific plugins. +Enterprise administrators can configure these controls on GitHub Enterprise Cloud for members with Copilot Business or Copilot Enterprise licenses issued by the enterprise or one of its organizations. +See managing plugin marketplaces and plugins with enterprise-managed settings for administrator setup. + +The post Control where GitHub Copilot customizations come from appeared first on The GitHub Blog.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New enterprise admin control over Copilot customization sources affects deployment, security posture, and plugin management strategy. |
| signals | new enterprise security control introduced · new managed setting: strictPluginOnlyCustomization · new policy scope: five customization categories (skills, custom agents, hooks, MCP servers, executable Copilot extensions) · authentication/authorization change: restricts where customizations originate · integration change: affects plugin marketplace and managed settings behavior · license scope specified: Copilot Business and Copilot Enterprise |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |