Merge branch 'main' into feature/pr-actions-diff-hunk

2026-08-29 03:32:08 +08:00 · 2026-08-11 13:08:52 -07:00
parent 58ca25f61f 64e5dd0363
commit b42d834865
19 changed files with 145 additions and 7 deletions
@@ -58,6 +58,7 @@ export const CollaborateNav: NavSection[] = [
        href: "/collaborate/enterprise/model-access-controls",
        children: "Model Access Controls",
      },
+      { href: "/collaborate/enterprise/groups", children: "Groups" },
      { href: "/collaborate/enterprise/audit-logs", children: "Audit Logs" },
      { href: "/collaborate/enterprise/migration", children: "Migration" },
    ],
@@ -114,6 +114,19 @@ A new git worktree is created from your current branch. The agent works in isola

 To create a worktree immediately from the default base branch, press `Cmd+Shift+N` (macOS) / `Ctrl+Shift+N` (Windows/Linux).

+### Slash Commands in the Worktree Prompt
+
+The new worktree prompt supports slash commands for its configuration options, so you can change them from the keyboard:
+
+| Command | Aliases | Action |
+|---|---|---|
+| `/models` | `/model` | Open the model selector |
+| `/agents` | `/modes` | Open the agent selector |
+| `/variant` | `/variants`, `/reasoning`, `/thinking` | Open the reasoning effort selector |
+| `/sandbox` | — | Toggle the sandbox for the new worktree |
+
+Navigate the menu with the arrow keys, select with `Enter` or `Tab`, and close it with `Escape`. Focus returns to the prompt after a selection or cancellation. `/agents` appears only when multiple agents are available, `/variant` only when the selected model has reasoning variants, and `/sandbox` only when sandbox controls are enabled.
+
 ### Multi-Version Mode

 You can run up to 4 parallel implementations of the same prompt across separate worktrees:
@@ -110,6 +110,7 @@ Type `@` in the TUI to get file autocomplete suggestions, or mention file paths
 - Messages → Expand/collapse details
 - Code blocks → Copy button appears
 - Mermaid code blocks → Fenced `mermaid` blocks render as diagrams after the message finishes streaming. The source remains copyable, and invalid Mermaid syntax stays visible in a contained error state.
+- Charts → In VS Code, asking for a chart, graph, or plot renders an inline chart in the conversation. Supported types include bar, line, scatter, pie, doughnut, radar, bubble, and polar area. Diagrams and flowcharts still render as Mermaid.

 **Status signals:**

@@ -43,6 +43,8 @@ Enable and sign in to the Kilo provider to use voice input in prompt fields. Req

 You can optionally choose a transcription model in **Settings** > **Models** > **Speech to Text Model**. Kilo stores this choice as `experimental.speech_to_text_model` in your global Kilo CLI config (`~/.config/kilo/kilo.jsonc`).

+The model list is discovered from the Kilo Gateway and reflects the transcription models available to your account or organization, so newly available models appear automatically.
+
 ---

 ## Record prompts
@@ -97,6 +97,8 @@ The `kilo console` command and its browser interface are deprecated and will be
 | `/copy` | - | Copy latest agent response |
 | `/copy-session` | - | Copy session transcript |
 | `/export` | - | Export session transcript |
+| `/resume-claude [uuid]` | - | Import a Claude Code session transcript |
+| `/resume-codex [uuid]` | - | Import an OpenAI Codex session transcript |
 | `/move` | - | Move the current session to another project directory |
 | `/diff` | - | Open the diff viewer |
 | `/timestamps` | `/toggle-timestamps` | Show/hide timestamps |
@@ -126,6 +128,7 @@ The `kilo console` command and its browser interface are deprecated and will be
 | `/reload` | - | Reload config, skills, agents, and commands from disk |
 | `/editor` | - | Open external editor |
 | `/auto-approve` | `/autoapprove`, `/approve-all`, `/approveall` | Toggle auto-approve mode for all permission prompts (saved to global config) |
+| `/privacy` | - | Toggle privacy mode (blurs PII in the TUI) |
 | `/exit` | `/quit`, `/q` | Exit the app |

 #### Kilo Gateway Commands (when connected)
@@ -143,6 +146,16 @@ The `kilo console` command and its browser interface are deprecated and will be
 | `/init` | Create/update AGENTS.md file for the project |
 | `/review` | Review code changes |

+### Importing Claude Code and Codex Sessions
+
+Continue work started in Claude Code or the OpenAI Codex CLI without copying transcripts by hand:
+
+1. Start a new session with `/new` — the import commands only run in an empty session.
+2. Run `/resume-claude` or `/resume-codex`.
+3. Pick one of the 10 most recent sessions for the current directory, or pass a session UUID directly, for example `/resume-claude <uuid>`.
+
+Kilo discovers Claude Code transcripts under `~/.claude/projects/` and Codex CLI rollouts under `~/.codex/sessions/`. The imported history keeps its original order and ends with an import notice. Content that Kilo cannot represent, such as some tool outputs, is skipped and counted in that notice.
+
 ## Local Code Reviews

 Review your code locally before pushing — catch issues early without waiting for PR reviews. Local code reviews give you AI-powered feedback on your changes without creating a public pull request.
@@ -389,6 +402,7 @@ Common configuration options include:
 - **`formatter`** - Code formatter configuration (`true`, `false`, or formatter-specific entries)
 - **`lsp`** - Language server configuration (`true`, `false`, or server-specific entries)
 - **`disabled_providers`** / **`enabled_providers`** - Control which providers are available
+- **`privacy_mode`** - Blur PII in the TUI (balance, team name, Kilo Pass usage) and require confirmation before `/profile` reveals account details — see [Privacy Mode](#privacy-mode)

 {% callout type="tip" %}
 **Using a model that's not in the built-in list?** You can register any model by adding it under `provider.<provider_id>.models` in your config file. See [Custom Models](/docs/code-with-ai/agents/custom-models) for full details and examples.
@@ -473,6 +487,23 @@ Kilo telemetry is enabled by default and can be disabled with `experimental.open

 If `OTEL_EXPORTER_OTLP_ENDPOINT` is set, the CLI exports OpenTelemetry traces and logs to that OTLP HTTP endpoint. You can also pass `OTEL_EXPORTER_OTLP_HEADERS` as comma-separated `key=value` pairs and `OTEL_RESOURCE_ATTRIBUTES` as comma-separated resource attributes. Request spans include `http.method`, `http.path`, route params such as `session.id` and `message.id`, and internal params under the `opencode.*` namespace.

+### Privacy Mode
+
+Set `privacy_mode` to `true` in `kilo.jsonc`, or toggle it with the `/privacy` command, to blur always-visible personal information in the TUI:
+
+```jsonc
+{
+  "privacy_mode": true,
+}
+```
+
+When privacy mode is on:
+
+- The sidebar footer shows the balance as `•••`, collapses the team name to "Team credits", and hides the Kilo Pass usage block.
+- `/profile` asks for confirmation before revealing your email, name, balance, and team on screen.
+
+Privacy mode only affects the TUI display. The `kilo profile` CLI command is unaffected.
+
 ### Environment Variables

 Use `{env:VARIABLE_NAME}` syntax in config files to reference environment variables:
@@ -522,6 +553,10 @@ Selecting an "Always run" option will:

 Kilo only saves the pattern you select. Approving a specific command does not approve redirected variants or broader command patterns unless that broader option is shown and selected.

+### Pasting Large Text
+
+Pasting a large block of text (five or more lines, or over 800 characters) into the prompt collapses it into a placeholder such as `[Pasted ~6 lines]` to keep the prompt readable. To view or edit the pasted text, paste the same text again — the matching placeholder expands in place.
+
 ## Autonomous Mode (Non-Interactive)

 Autonomous mode allows Kilo Code to run in automated environments like CI/CD pipelines without requiring user interaction.
@@ -125,6 +125,7 @@ When remote mode is enabled in the CLI, your active local sessions appear in the
 - **Permission requests** route to your active connection
 - **Full editing capabilities** work remotely
 - **Session renames** sync in both directions between the CLI and the web or mobile app
+- **File attachments** sync with the mobile app — send files from your phone to the CLI, and receive files the agent delivers back. See [Attachments in remote sessions](/docs/code-with-ai/platforms/mobile#attachments-in-remote-sessions)

 ### Enabling Remote Mode

@@ -27,6 +27,12 @@ The mobile app lets you:
 - Review GitHub pull requests end to end — diffs, checks, comments, and merging.
 - Start a new session on a connected `kilo remote` CLI instance with the **Run on** picker.

+## Privacy and telemetry
+
+On first launch, the app asks for your consent before enabling optional telemetry — product analytics, attribution, and performance tracing. Optional telemetry is pre-selected during onboarding; you can turn it off before accepting. No optional analytics starts before you make a choice.
+
+You can review or change your decision at any time in **Settings**. Declining optional telemetry keeps you signed in, and optional telemetry state stored on the device is stopped and discarded when you revoke consent or sign out.
+
 ## Kilo Pass and Billing

 For Kilo Pass pricing, billing, and account management details, use the [Kilo Pass pricing page](https://kilo.ai/pricing/kilo-pass).
@@ -29,6 +29,7 @@ Key features include:
 - **[Workflows](/docs/customize/workflows)** — Repeatable prompt templates as `.md` files
 - **[Skills](/docs/customize/skills)** — Load specialized domain knowledge from SKILL.md files
 - **[Custom Subagents](/docs/customize/custom-subagents)** — Define specialized sub-agents for the `task` tool
+- **Charts** — Ask for a chart or plot and get an inline data visualization in the chat
 - **Open in Tab** — Pop the chat out into a full editor tab
 - **Transcript export:** Save complete local session transcripts as Markdown files
 - **Sub-Agent Viewer** — Read-only panels for viewing child agent sessions
@@ -0,0 +1,56 @@
+---
+title: "Groups"
+description: "Organize members into groups and compose policies like model access"
+---
+
+# Groups
+
+{% callout type="info" %}
+This is an **Enterprise-only** feature. The Groups page and policy enforcement are available only to organizations on the Enterprise plan.
+{% /callout %}
+
+**Groups** let you organize members of your organization and attach **policies** to them, without creating sub-organizations. Groups are flat (no nesting) and a member can belong to any number of groups.
+
+**Model access** is the first policy type. It grants models and providers to the members of a group, layered on top of the organization-wide [Model Access Controls](/docs/collaborate/enterprise/model-access-controls). More policy types are planned.
+
+## Roles
+
+| Role | What they can do |
+|---|---|
+| Owner | Create, edit, and delete groups; manage members and policies |
+| Billing manager | View groups and policy settings (read-only) |
+| Member | See the names of the groups they belong to |
+
+## Managing groups
+
+Open **Groups** in your organization's sidebar on the [Organization dashboard](https://app.kilo.ai).
+
+1. Click **Create group**
+2. Give the group a name and optional description
+3. Add members
+4. Add policies
+
+Deleting a group removes the group, its policies, and its member assignments. Members keep whatever access the organization defaults and their remaining groups grant.
+
+## Model access policies
+
+A model access policy has one of three modes:
+
+| Mode | Effect |
+|---|---|
+| **All** | Grants every model and provider |
+| **Selected** | Grants only the models and providers you select |
+| **None** | Grants nothing |
+
+Policies compose in layers:
+
+- **Default policies** apply to every direct member before their group policies are combined. Manage them from the **Group policies** card on the Groups page.
+- **Group policies** apply to the group's members. A member in several groups gets the union of those grants.
+
+A member's effective access is the organization ceiling intersected with the default policies plus the union of their group grants:
+
+- Absence of configuration is never a restriction — if no default or group policy applies to a member, their access is unchanged.
+- Group grants cannot exceed the organization-wide [Model Access Controls](/docs/collaborate/enterprise/model-access-controls). Models and providers blocked there stay blocked for everyone, and the policy editor marks out-of-ceiling entries as unavailable.
+- Only an explicit **None** mode, or a **Selected** mode with nothing selected, results in no access.
+
+Policy changes are enforced everywhere members use models, including the extension, the CLI, autocomplete, and the Slack, Discord, Linear, and GitHub integrations.
@@ -75,3 +75,4 @@ Use filters to find the models or providers you want to block:
 - Individual users cannot override organization-level restrictions.
 - Blocking a provider blocks all its models, including models added by that provider in the future.
 - Unblocking a provider immediately restores access to all its models.
+- To grant models to specific sets of members instead of the whole organization, use [Groups](/docs/collaborate/enterprise/groups). These organization-wide controls remain a hard ceiling that group grants cannot exceed.
@@ -66,6 +66,18 @@ To fill empty seats or remove members ahead of seat deletion, see the [team mana

 To fill empty seats or remove members ahead of seat deletion, see the [team management](/docs/collaborate/teams/team-management) page.

+## Kilo Pass for Organizations
+
+Organizations can subscribe to [Kilo Pass](https://kilo.ai/pricing/kilo-pass) at the organization level, pooling the subscription's credit capacity across the team instead of each member holding a personal Kilo Pass. Personal Kilo Pass subscriptions stay separate and are not affected.
+
+- **One pass per seat** — purchased capacity always matches your paid seat count and adjusts automatically as seats are added or removed.
+- **Pooled credits** — each service window's Kilo Pass credits are issued to the organization's pool rather than to individuals.
+- **Child organizations** — if your organization has direct child organizations, you can allocate part of the pooled capacity to each child. Anything not allocated stays with the parent organization.
+
+Organization owners and billing managers can purchase Kilo Pass, manage allocations, and cancel from your organization's **Subscriptions** page on the [Organization dashboard](https://app.kilo.ai). Cancellation takes effect at the end of the current paid period. Regular members do not see these controls.
+
+If a seat reduction leaves your allocations larger than the remaining capacity, the subscription becomes overallocated and the Subscriptions page walks you through reconciling it.
+
 ## Automatic Top-Up

 Ensure your team has uninterrupted access to Kilo Code by enabling [Automatic Top-Up](https://kilo.ai/features/auto-top-ups). This feature keeps your organization's balance funded so you never have to worry about manual recharges. Auto top-up is also available for individual accounts — see [Adding Credits](/docs/getting-started/adding-credits) for details.
@@ -132,7 +132,7 @@ flowchart LR

 ## Identity and access

-Web control plane uses JWT-backed application sessions and supports multiple sign-in methods. Repository-supported providers include Google, Apple, GitHub, GitLab, Discord, LinkedIn OpenID Connect, WorkOS enterprise SSO, and email magic links.
+Web control plane uses JWT-backed application sessions and supports multiple sign-in methods. Repository-supported providers include Google, Apple, GitHub, GitLab, Discord, LinkedIn OpenID Connect, Anaconda, WorkOS enterprise SSO, and email magic links.

 Kilo Cloud uses several authorization contexts:

@@ -57,8 +57,15 @@ You are helping submit a pull request...
 | `description` | Shown in the command picker |
 | `agent` | Which agent to use when invoking this command |
 | `model` | Model override for this command |
+| `variant` | Reasoning effort variant override (for example `low` or `high`), for models that support variants |
 | `subtask` | When `true`, runs as a sub-agent session |

+### Model and Reasoning Variant
+
+Each workflow can run with its own model and reasoning effort variant. In the VS Code extension, open **Settings → Agent Behaviour → Workflows**, expand a workflow, and choose a model and variant. The selection is saved as a command override in your global config, so the workflow's template file stays unchanged.
+
+You can also set `model` and `variant` in the command's frontmatter or in the `command` section of `kilo.jsonc`.  A variant only applies when the selected model supports it — picking a different model clears a variant the new model does not offer.
+
 ### Workflow Capabilities

 Workflows can leverage all built-in tools: `read`, `glob`, `grep`, `edit`, `write`, `bash`, `webfetch`, and MCP server tools.
@@ -87,7 +87,7 @@ BYOK lets you use your own provider API keys with the Kilo AI Gateway. When a BY
 | OpenAI | `openai` |
 | MiniMax | `minimax` |
 | Mistral | `mistral` |
-| xAI | `xai` |
+| SpaceXAI | `xai` |
 | Z.AI | `zai` |
 | BytePlus Coding Plan | `byteplus-coding` |
 | Chutes BYOK | `chutes-byok` |
@@ -11,7 +11,7 @@ The gateway works seamlessly with the [Vercel AI SDK](https://ai-sdk.dev), the [

 ## Key features

- **One key, hundreds of models**: Access models from Anthropic, OpenAI, Google, xAI, Mistral, MiniMax, and more with a single API key
+- **One key, hundreds of models**: Access models from Anthropic, OpenAI, Google, SpaceXAI, Mistral, MiniMax, and more with a single API key
 - **OpenAI-compatible API**: Drop-in replacement for OpenAI's `/chat/completions` endpoint -- switch models by changing a single string
 - **Streaming support**: Full Server-Sent Events (SSE) streaming with time-to-first-token tracking
 - **BYOK (Bring Your Own Key)**: Use your own provider API keys with encrypted-at-rest storage
@@ -48,8 +48,8 @@ This returns model information including pricing, context window, and supported
 | `openai/gpt-5.4-mini` | OpenAI | Fast and efficient |
 | `google/gemini-3.1-pro-preview` | Google | Advanced reasoning |
 | `google/gemini-2.5-flash` | Google | Fast and efficient |
-| `x-ai/grok-4` | xAI | Most capable Grok model |
-| `x-ai/grok-code-fast-1` | xAI | Optimized for code tasks |
+| `x-ai/grok-4` | SpaceXAI | Most capable Grok model |
+| `x-ai/grok-code-fast-1` | SpaceXAI | Optimized for code tasks |
 | `deepseek/deepseek-v3.2` | DeepSeek | Strong coding and reasoning model |
 | `moonshotai/kimi-k2.5` | Moonshot | Strong coding and multilingual model |
 | `minimax/minimax-m2.7` | MiniMax | High-performance MoE model |
@@ -63,6 +63,7 @@ Several models are available at no cost, subject to rate limits:
 | `stepfun/step-3.7-flash:free` | StepFun Step 3.7 Flash |
 | `poolside/laguna-m.1:free` | Poolside Laguna M.1 |
 | `nvidia/nemotron-3-ultra-550b-a55b:free` | NVIDIA Nemotron 3 Ultra |
+| `tencent/hy3:free` | Tencent Hy3 |
 | `openrouter/free` | Best available free model |

 Free models are available to both authenticated and anonymous users. Anonymous users are rate-limited to 200 requests per hour per IP address.
@@ -34,7 +34,7 @@ Use your provider API key to route matching models through your account:
 - Novita
 - OpenAI
 - Xiaomi
- xAI
+- SpaceXAI
 - Z.ai

 ### Subscription and direct provider plans
@@ -5,7 +5,7 @@ description: "Configure Kilo Code and connect to your AI providers"

 # Authentication

-When you install Kilo Code, you'll be prompted to sign in or create a free account. This automatically configures everything you need to get started.
+When you install Kilo Code, you'll be prompted to sign in or create a free account. This automatically configures everything you need to get started. You can sign in with your email address, Google, Apple, GitHub, GitLab, Discord, LinkedIn, or Anaconda, or through your organization's single sign-on.

 ## Quick Start with Kilo Account

@@ -31,6 +31,7 @@ Depending on your setup, you can also use:

 ## Creating an Instance

+
 1. Navigate to your [Kilo profile](https://app.kilo.ai/profile)
 2. Click **Claw** in the left navigation