Merge pull request #45760 from github/repo-sync

Repo sync

Author: docs-bot

content/copilot/concepts/about-cloud-and-local-sandboxes.md

@@ -53,7 +53,7 @@ To enable local sandboxing inside a {% data variables.copilot.copilot_cli_short
 /sandbox enable
 ```
 
-After you enable local sandboxing, the commands and tools that an agent runs on your behalf—shell commands, file search, and, by default, the MCP and language (LSP) servers the CLI starts—run inside an operating-system-level sandbox, limiting their access to your system. The CLI continues to use local sandboxing whenever you use the CLI in future—for programmatic as well as interactive use—until you run `/sandbox disable` to disable it. If enterprise managed settings require sandboxing, you cannot disable it.
+After you enable local sandboxing, the commands and tools that an agent runs on your behalf—shell commands, file search, and, by default, the MCP and language (LSP) servers the CLI starts—run inside an operating-system-level sandbox, limiting their access to your system. The CLI continues to use local sandboxing whenever you use the CLI in future—for programmatic as well as interactive use—until you run `/sandbox disable` to disable it. If enterprise managed settings require sandboxing, ordinary settings, startup options, and `/sandbox disable` cannot turn it off. If the effective policy permits sandbox bypass, you can still explicitly disable sandboxing for the rest of the current session from an active bypass permission prompt.
 
 The CLI's built-in file tools—first-party commands that are part of the CLI, rather than shell commands like `sed`—run in-process in the CLI. Because the CLI itself is not sandboxed, the operating-system sandbox never sees the file operations these tools perform and cannot constrain them. Instead, the built-in tools are coded to check the sandbox policy themselves and honor your configured settings on a best-effort basis.
 

content/copilot/concepts/agents/cloud-agent/about-custom-agents.md

@@ -23,6 +23,7 @@ Once you create {% data variables.copilot.custom_agents_short %}, they become av
 
 * **{% data variables.copilot.copilot_cloud_agent %} on {% data variables.product.prodname_dotcom_the_website %}**: The agents tab and panel, issue assignment, and pull requests
 * **{% data variables.copilot.copilot_cloud_agent %} in IDEs**: {% data variables.product.prodname_vscode %}, JetBrains IDEs, Eclipse, and Xcode
+* **{% data variables.copilot.github_copilot_app %}**
 * **{% data variables.copilot.copilot_cli %}**
 
 You can use {% data variables.copilot.agent_profiles %} directly in {% data variables.product.prodname_vscode %}, JetBrains IDEs, Eclipse, and Xcode. Some properties may function differently or be ignored between environments.

content/copilot/concepts/agents/copilot-cli/understanding-local-sandboxing.md

@@ -111,7 +111,7 @@ You can grant extra read/write or read-only paths, deny paths, and change other
 
 ## Enterprise-managed policies
 
-If you get {% data variables.product.prodname_copilot_short %} through an enterprise-owned organization, an administrator can enforce a filesystem policy through managed settings. Managed settings act as a baseline that you cannot loosen: they can require sandboxing, add denied paths, and limit which paths you are allowed to grant. Where a managed setting applies, the `/sandbox config` dialog shows it as a locked **(managed)** value, and `/sandbox policy` reflects it in the resolved policy.
+If you get {% data variables.product.prodname_copilot_short %} through an enterprise-owned organization, an administrator can enforce a filesystem policy through managed settings. Managed settings act as a restrictive baseline: they can require sandboxing, add denied paths, and limit which paths you are allowed to grant. Where a managed setting applies, the `/sandbox config` dialog shows it as a locked **(managed)** value, and `/sandbox policy` reflects it in the resolved policy. If the effective policy permits sandbox bypass, a user can explicitly disable sandboxing for the rest of the current session from an active bypass permission prompt. This session opt-out does not loosen the saved policy.
 
 Unlike most settings, where a single source wins, the sandbox policy is composed from every source in force at once. Managed settings can arrive through more than one channel simultaneously—server-managed, MDM, and file-based—and these combine with each other, and with your own settings, in the **most restrictive** direction rather than one source overriding another: a required toggle stays on, denied paths from all sources add up, and the paths you are allowed to grant can only be narrowed. For more information, see [AUTOTITLE](/copilot/reference/enterprise-administrators/enterprise-managed-settings#sandbox).
 

content/copilot/how-tos/cloud-and-local-sandboxes/configuring-local-sandbox-settings.md

@@ -36,7 +36,7 @@ The **General** tab controls the top-level sandbox behavior. When enterprise man
 | Setting | Description |
 | --- | --- |
 | **Enable sandbox** | Run shell commands inside the sandbox. You can also toggle this with `/sandbox enable` and `/sandbox disable`. |
-| **Allow sandbox bypass** | Let the model request that individual commands run outside the sandbox, subject to approval. Turned on by default. For more information, see [Allowing sandbox bypass](#allowing-sandbox-bypass). |
+| **Allow sandbox bypass** | Let the model request that individual commands run outside the sandbox, subject to approval. A bypass prompt can also let you disable sandboxing for the rest of the current session. Turned on by default. For more information, see [Allowing sandbox bypass](#allowing-sandbox-bypass). |
 | **Sandbox MCP servers** | Run MCP servers inside the sandbox. Turned on by default. |
 | **Sandbox LSP servers** | Run language servers (LSP servers) inside the sandbox. Turned on by default. |
 
@@ -47,6 +47,8 @@ The **Allow sandbox bypass** setting controls what happens when {% data variable
 * **On (default)**: If a command fails inside the sandbox, you are prompted to allow {% data variables.product.prodname_copilot_short %} to run the command outside the sandbox. Your response to this prompt applies to this specific attempt to run the command. Optionally, you can choose to disable the sandbox for the rest of the session (if permitted by your enterprise), or you can enter an instruction for {% data variables.product.prodname_copilot_short %} to work on instead.
 * **Off**: If {% data variables.product.prodname_copilot_short %} can't run a command successfully in the sandbox, it stops working on the task and reports the failure.
 
+If enterprise managed settings set `sandbox.allowBypass` to `false`, you cannot approve individual commands to run outside the sandbox or disable sandboxing for the rest of the session. If managed settings require sandboxing but the effective policy permits bypass, you can disable sandboxing only from an active bypass permission prompt, not through ordinary settings or `/sandbox disable`.
+
 ## Configuring authentication settings
 
 The **Auth** tab controls whether your credentials are made available to commands running inside the sandbox. As on the other tabs, an enterprise-managed value is shown as `(managed)` and can't be changed.

content/copilot/how-tos/cloud-and-local-sandboxes/using-local-sandboxing.md

@@ -20,7 +20,7 @@ docsTeamMetrics:
 
 Sandboxing is currently an experimental feature. To use it, start {% data variables.copilot.copilot_cli_short %} with the `‑‑experimental` command line option, or enter `/experimental on` during a session.
 
-When you enable local sandboxing, {% data variables.copilot.copilot_cli_short %} runs most of the commands and tools it invokes on your behalf inside an operating-system sandbox. After you enable local sandboxing, it is used for all your {% data variables.copilot.copilot_cli_short %} sessions until you disable it, or turn it off for a specific session. If enterprise managed settings require sandboxing, you cannot disable it.
+When you enable local sandboxing, {% data variables.copilot.copilot_cli_short %} runs most of the commands and tools it invokes on your behalf inside an operating-system sandbox. After you enable local sandboxing, it is used for all your {% data variables.copilot.copilot_cli_short %} sessions until you disable it, or turn it off for a specific session. If enterprise managed settings require sandboxing, ordinary configuration, the `--no-sandbox` command line option, and the `/sandbox disable` command cannot disable it. However, if the effective policy permits sandbox bypass, you can explicitly disable sandboxing for the rest of the current session from an active bypass permission prompt.
 
 By default, sandboxed commands and tools can write within your current working directory and temporary folders. Your user profile (home) directory, along with system and tool locations are read-only. Other disk locations are blocked. In a Git repository, the rest of the repository above your current working directory is readable but not writable. Access to your local and private network is permitted, as is outbound internet access.
 
@@ -60,7 +60,7 @@ After you enable local sandboxing, it continues to be used for the current and f
 
 ## Disabling local sandboxing
 
-If enterprise managed settings require sandboxing, you cannot disable it, and `/sandbox disable` is refused.
+If enterprise managed settings require sandboxing, `/sandbox disable` is refused. If the effective policy permits sandbox bypass, you can instead explicitly disable sandboxing for the rest of the current session from an active bypass permission prompt.
 
 To stop using local sandboxing, enter the following command in an interactive {% data variables.copilot.copilot_cli_short %} session:
 
@@ -84,7 +84,7 @@ copilot --sandbox -p "PROMPT"
 
 ## Running a single command outside the sandbox
 
-When a command needs broader access than the sandbox allows, {% data variables.product.prodname_copilot_short %} can request to run that single command outside the sandbox. You are shown a confirmation prompt describing the command, and it runs outside the sandbox only if you approve it; otherwise it stays sandboxed. The rest of your session remains sandboxed either way.
+When a command needs broader access than the sandbox allows, {% data variables.product.prodname_copilot_short %} can request to run that single command outside the sandbox. You are shown a confirmation prompt describing the command. You can approve that single command, keep it inside the sandbox, or disable sandboxing for the rest of the current session. The session opt-out is available only while responding to an active bypass prompt and only if the effective policy permits sandbox bypass.
 
 This behavior is enabled by default and can be turned off in your sandbox settings.
 

content/copilot/reference/ai-models/model-hosting.md

@@ -27,6 +27,7 @@ Used for:
 * {% data variables.copilot.copilot_gpt_56_luna %}
 * {% data variables.copilot.copilot_gpt_56_sol %}
 * {% data variables.copilot.copilot_gpt_56_terra %}
+* {% data variables.copilot.copilot_gpt_6_astra %}
 
 These models are hosted by OpenAI and {% data variables.product.github %}'s Azure infrastructure.
 

content/copilot/reference/ai-models/supported-models.md

@@ -92,6 +92,7 @@ Choosing a larger context window or higher reasoning will impact {% data variabl
 | {% data variables.copilot.copilot_gpt_56_luna %} | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} |
 | {% data variables.copilot.copilot_gpt_56_sol %} | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} |
 | {% data variables.copilot.copilot_gpt_56_terra %} | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} |
+| {% data variables.copilot.copilot_gpt_6_astra %} | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} |
 | {% data variables.copilot.copilot_kimi_k3 %} | {% octicon "check" aria-label="Supported" %} | {% octicon "check" aria-label="Supported" %} |
 
 {% endrowheaders %}
@@ -133,6 +134,7 @@ Some {% data variables.product.prodname_copilot_short %} models require minimum
 | {% data variables.copilot.copilot_gpt_56_luna %}         | `1.128.0` | TBD | TBD | TBD | TBD |
 | {% data variables.copilot.copilot_gpt_56_sol %}          | `1.128.0` | TBD | TBD | TBD | TBD |
 | {% data variables.copilot.copilot_gpt_56_terra %}        | `1.128.0` | TBD | TBD | TBD | TBD |
+| {% data variables.copilot.copilot_gpt_6_astra %}         | `1.136.1` | `17.14.19` | TBD | TBD | TBD |
 | {% data variables.copilot.copilot_claude_opus_48 %}      | `v1.118` | `17.14.6`                     | TBD | TBD | TBD |
 | {% data variables.copilot.copilot_claude_opus_5 %}       | `v1.128.0` | `17.14.22` | TBD | TBD | TBD |
 | {% data variables.copilot.copilot_claude_sonnet_5 %}  | `v1.124` | `17.14.6` | TBD | TBD | TBD |

content/copilot/reference/copilot-billing/models-and-pricing.md

@@ -33,7 +33,7 @@ All prices are **per 1 million tokens**.
 
 {% data reusables.copilot.extended-context-pricing %}
 
-{% data variables.copilot.copilot_gpt_56_sol %}, {% data variables.copilot.copilot_gpt_56_terra %}, and {% data variables.copilot.copilot_gpt_56_luna %} include a cache write cost in addition to cached input. Earlier OpenAI models have no cache write cost.
+{% data variables.copilot.copilot_gpt_56_sol %}, {% data variables.copilot.copilot_gpt_56_terra %}, {% data variables.copilot.copilot_gpt_56_luna %}, and {% data variables.copilot.copilot_gpt_6_astra %} include a cache write cost in addition to cached input. Earlier OpenAI models have no cache write cost.
 
 | Model | Release status | Category | Tier | Threshold (input tokens) | Input | Cached input | Cache write | Output |
 | --- | --- | --- | --- | --- | ---: | ---: | ---: | ---: |

content/copilot/reference/copilot-cli-reference/cli-command-reference.md

@@ -609,11 +609,11 @@ Plan-then-autopilot lets a session start in plan mode and automatically continue
 
 ### Enterprise-managed sandbox floor
 
-An enterprise-managed policy can enforce OS-level shell sandboxing as a minimum floor. In other words, even if you pass `--no-sandbox`, the policy can still force sandboxing on. This is a policy override, not a failure of the flag itself. By contrast, `--sandbox` is unaffected because it only turns sandboxing on and never removes it.
+An enterprise-managed policy can enforce OS-level shell sandboxing as a minimum floor. In other words, even if you pass `--no-sandbox`, the policy can still force sandboxing on. This is a policy override, not a failure of the flag itself. By contrast, `--sandbox` is unaffected because it only turns sandboxing on and never removes it. If the effective policy permits sandbox bypass, you can explicitly disable sandboxing for the rest of the current session while responding to an active bypass permission prompt.
 
 When a managed policy overrides your setting, the CLI shows a warning in the interactive timeline (or on stderr when using `-p`) so it is clear that the behavior comes from policy enforcement rather than the option failing to work. Contact your administrator if you need the policy changed. The `/sandbox` command is also registered whenever a managed policy forces sandboxing on, even without experimental features enabled, so you can still inspect the effective policy and status while the floor applies. {% data reusables.copilot.experimental %}
 
-Adding the managed `sandbox.failIfUnavailable` setting set to `true`, alongside `sandbox.enabled` set to `true`, makes the sandbox mandatory. Instead of falling back to running commands unsandboxed, {% data variables.product.prodname_copilot_short %} blocks model and tool execution if the policy can't be validated, compiled, or enforced by a usable sandbox backend, and you can't disable it. See [AUTOTITLE](/copilot/reference/copilot-cli-reference/cli-config-dir-reference#user-settings-copilotsettingsjson).
+Adding the managed `sandbox.failIfUnavailable` setting set to `true`, alongside `sandbox.enabled` set to `true`, makes the sandbox mandatory when it cannot be established. Instead of falling back to running commands unsandboxed, {% data variables.product.prodname_copilot_short %} blocks model and tool execution if the policy can't be validated, compiled, or enforced by a usable sandbox backend. See [AUTOTITLE](/copilot/reference/enterprise-administrators/enterprise-managed-settings#sandbox).
 
 The CLI also warns when a managed policy enables sandboxing in a session that you did not request, not only when it overrides `--no-sandbox`. This includes sessions where the policy arrives after startup, because server-managed settings are only available after login. The warning is omitted if your own settings or the `--sandbox` option already requested sandboxing, since the session state would then be expected.