diff --git a/packages/kilo-docs/pages/automate/agent-manager-workflows.md b/packages/kilo-docs/pages/automate/agent-manager-workflows.md index 41db1c0f593a..e91e5d5e008c 100644 --- a/packages/kilo-docs/pages/automate/agent-manager-workflows.md +++ b/packages/kilo-docs/pages/automate/agent-manager-workflows.md @@ -169,6 +169,7 @@ Layer review in before asking a teammate: - **`/review`** — slash command, AI review of staged, unstaged, and untracked changes in the worktree when run without arguments. Good as a last pass before committing. - **`/review uncommitted [guidance]`** — explicitly review uncommitted changes, optionally focusing the review with guidance. - **`/review branch [base] [guidance]`** — review the whole branch vs. its detected or specified base, with optional guidance. +- **`/review worktree [guidance]`** - review committed, staged, unstaged, and untracked changes against the worktree's recorded parent branch. Available only in Agent Manager managed worktree sessions. - **`/review ` or `/review `** — review a specific commit or pull request. - **`kilo review` in CI** — automated PR review. See [Code Reviews](/docs/automate/code-reviews/overview) for the setup. - **Human review** — push the branch from the session terminal and `gh pr create`. The PR badge appears on the worktree and stays in sync with CI and reviews. Review, comment on, and merge the pull request from the internal PR panel; see [Reviewing a pull request](/docs/automate/agent-manager#reviewing-a-pull-request). @@ -198,11 +199,7 @@ Three ways, pick based on how much collaboration the change needs: ### Parent branch → worktree -When the parent branch moves ahead, ask the agent from the worktree's session: - -> Merge the latest `origin/main` into this branch and resolve any conflicts. Do not use `git stash`. - -Save this as a reusable slash command if you do it often. +When the parent branch moves ahead, run `/update-from-base` in the managed worktree's chat. It asks the agent to fetch and merge the saved base, preserving uncommitted edits without Git stash. The [Push Pull Request Fixes](/docs/automate/agent-manager#push-pull-request-fixes) setting controls whether it is also asked to push after checks pass. See [Update from the base branch](/docs/automate/agent-manager#update-from-the-base-branch) for details. {% callout type="danger" %} **Never use `git stash` inside a worktree.** Stashes live in the shared `.git` directory that every worktree points at, so a stash made in one worktree can be popped in another — crossing uncommitted changes between agents. Use a WIP commit or a temporary branch instead. @@ -216,7 +213,7 @@ The Agent Manager is good at conflict resolution when you give it context. A low ### When several worktrees finish at once -Merge the most foundational one first. Then, in each remaining worktree, ask the agent to pull the updated parent branch in (same prompt as above) before merging. The agent handles the merge direction and only escalates conflicts it cannot resolve. +Merge the most foundational one first. Then run `/update-from-base` in each remaining worktree before merging it. Give the agent context when a conflict needs a decision about the intended behavior. ## Hygiene diff --git a/packages/kilo-docs/pages/automate/agent-manager.md b/packages/kilo-docs/pages/automate/agent-manager.md index a110d1275faf..9ca6512f8152 100644 --- a/packages/kilo-docs/pages/automate/agent-manager.md +++ b/packages/kilo-docs/pages/automate/agent-manager.md @@ -103,6 +103,15 @@ Agent Manager worktree defaults belong to a repository. Open a project's setting The **Worktree Setup Script** control opens or creates the setup script for the selected repository. See [Setup Scripts](#setup-scripts) for supported filenames and execution behavior. +### Branch naming + +The same settings tab has two application-wide controls: + +- **Automatic branch naming** is on by default. Kilo names a branch once the conversation describes a clear task. It does not rename explicitly named or published branches. +- **Branch prefix** adds a prefix such as `feature/` to automatically named branches. It is empty by default and does not affect explicit names. + +Use **Save** to apply these settings. + ## Providers and Authentication Agent Manager uses the same sign-in, provider settings, models, BYOK keys, custom providers, MCP servers, and permission rules as the extension sidebar. Configure them from extension Settings and they apply to Agent Manager as well. @@ -119,7 +128,11 @@ In a managed worktree's chat, type `/update-from-base` and select the action to The saved base stays the same if you switch branches in Local or change the project's default base. For example, a worktree created from `main` still updates from `main` when Local has `release` checked out. If you switch branches inside the managed worktree, the agent updates that worktree's current branch, not its original branch. Select the intended worktree before running the command; it does not update Local. -The agent uses the recorded remote, or the saved base branch's upstream if no remote was recorded. It asks for a source if the base is local-only or unavailable. The request prohibits stashing, discarding uncommitted work, and pushing. Existing merge or rebase operations and blocking dirty changes require your input. Normal tool approvals still apply. +The agent uses the recorded remote, or the saved base branch's upstream if no remote was recorded. It asks for a source if the base is local-only or unavailable. + +The request instructs the agent to preserve staged, unstaged, and untracked edits in a verified recovery copy for this worktree, merge the freshly fetched base, then restore the edits and their staging state. It must not use Git stash or include unfinished work in the merge commit. If it cannot verify preservation, a merge or rebase is already in progress, or a conflict needs your decision, it stops and asks. + +When [Push Pull Request Fixes](#push-pull-request-fixes) is on and the worktree has a pull request, the request also asks the agent to push after a clean merge and successful checks. Otherwise it tells the agent not to push. Normal tool permissions still apply. ### Worktree Location @@ -218,7 +231,7 @@ The summary shows the merge state before the merge controls: Merge controls appear only when you have write permission. Choose **Create merge commit**, **Squash and merge**, or **Rebase and merge** from the split button. Agent Manager remembers the last method you used per repository and preselects it when the repository allows it. - **Update branch:** shown when the branch is behind. This asks GitHub to merge the base branch into the PR branch. -- **Fix with Kilo:** shown for merge conflicts. It lists the conflicting files and starts an update-from-base request, which asks the agent to resolve the conflicts in the worktree. +- **Fix with Kilo:** shown for merge conflicts. It asks the agent to fetch and merge the worktree's saved base branch and resolve conflicts, using the [update-from-base workflow](#update-from-the-base-branch). - **Enable auto-merge:** available when the repository allows auto-merge. GitHub merges the pull request when its requirements are met. Use **Disable auto-merge** to cancel it. - **Merge pull request:** when the branch is clean, the primary button merges immediately after a confirmation that names the selected method. @@ -249,7 +262,7 @@ Sending a thread gives it to Kilo as review context, and it does not post anythi #### Checks -Each failing check offers **Fix with Kilo**, which sends the failure summary and log commands to the current agent. With an active Agent Manager terminal, the label becomes **Send failures to terminal**. +When checks fail or are cancelled, the Checks section offers **Fix with Kilo**, which sends a failure summary and log commands to the current agent. With an active Agent Manager terminal, the label becomes **Send failures to terminal**. Sending this context does not post a GitHub comment; the [push-fixes setting](#push-pull-request-fixes) controls whether the agent is also asked to commit and push its fix. #### Conversation @@ -263,6 +276,8 @@ The conversation lists the pull request description and history in one timeline: Use a comment card's **Fix with Kilo** action to hand it to the agent. **Dismiss** hides a comment from the next send, and **Restore** brings it back. **Fix N with Kilo** or **Send N to terminal** sends the actionable comments together. +Batch sends skip bot comments, dismissed comments, and comments already sent. Use **Add comment** to post a new discussion comment to GitHub, or **Copy comment** to copy an existing one as Markdown. + #### Reviewers The Reviewers section shows each requested or completed reviewer with an avatar and one of these GitHub-style states: **Approved**, **Changes requested**, **Commented**, or **Awaiting**. @@ -274,8 +289,10 @@ The Reviewers section shows each requested or completed reviewer with an avatar - Set it in **Settings > Agent Behaviour > Push Pull Request Fixes** (`kilo-code.new.agentManager.pushFixes`). - Toggle the same setting from the PR panel header with the **Push Pull Request Fixes** button. - When it is on, fix prompts for PR review comments and CI failures include: "When the changes pass local checks, commit them and push to this branch so the pull request updates. Do not force-push." -- Update from base follows the same setting. With it on, the agent is asked to push the branch after a clean merge. With it off, it is not. -- Permission prompts still confirm commits and pushes. Turning the setting on does not auto-approve them. +- Update from base follows the same setting: when the worktree has a pull request, the agent is asked to push after a clean merge and successful checks. +- Your configured tool permissions still apply. Auto-approved commands may run without another prompt. When sandboxing is active, Git commands that change repository state require separate, one-shot approval. + +Turn the setting off to keep fixes local for manual commit and push. Local draft comments sent from the diff panel do not include the push instruction. ### Creating a New Worktree Session @@ -285,6 +302,8 @@ The Reviewers section shows each requested or completed reviewer with an avatar Kilo creates the worktree from the selected project's configured default base branch. In a multi-project workspace, the selected project determines this setting. An explicit base branch selected in the dialog takes precedence. If no default is configured, Kilo falls back to automatic detection of the repository's remote default branch. The agent works in isolation, so your main branch is unaffected. +Explicit branch names preserve slashes and case, such as `feature/MyTask`. The checkout directory can have a different name. + To create a worktree immediately from the default base branch, press `Cmd+Shift+N` (macOS) / `Ctrl+Shift+N` (Windows/Linux). This uses the selected project's configured default, or the automatic remote-default fallback when no configured default exists. ### Slash Commands in the Worktree Prompt @@ -322,7 +341,8 @@ Imported work stays associated with its branch or worktree and can be continued - Create a worktree session to start a new agent in an isolated branch - Press `Cmd+T` (macOS) / `Ctrl+T` (Windows/Linux) to start another session in the selected worktree - Sessions in one worktree have separate transcripts and prompt queues, but share the same checkout, branch, and terminal state. Coordinate write-heavy work before asking multiple sessions to edit the same files. -- Use session history to reopen local sessions or preview cloud sessions +- Use the history button in the **WORKTREES** header to reopen local sessions or preview cloud sessions. In multi-project mode, each project's history button opens that project's Local history. +- Hover a session in history for actions to open it in a new worktree or move it to the project's Local tabs. - When a worktree is selected, open session history to use the **Worktree** source, which is selected by default and lists only sessions assigned to that worktree. Opening a worktree session returns to its owning worktree. - Continue a cloud session locally from Agent Manager using the same extension sign-in and provider settings @@ -347,15 +367,17 @@ The tool supports two modes: | `worktree` | Creates one Agent Manager git worktree and session per task. Each task gets an isolated checkout, branch, directory, and terminal. | | `local` | Creates Agent Manager sessions in the current workspace without git worktree isolation. You can also target an existing managed worktree with its `worktreeID`. | -Each request can include 1-20 tasks. Each task must include at least one of `prompt`, `name`, or `branchName`. Prompted tasks inherit the model and reasoning variant used by the chat turn that starts them. A task can override that selection with a `model` (by name, e.g. `Claude Opus 4.1`) when you explicitly request a different model, or with one of the current model's reasoning `variant` values when you request a different variant. Add `provider` beside `model` to force a model-name match to one of the listed provider IDs. Agent Manager resolves the provider for a model override when `provider` is omitted, preferring the provider used by the current turn and falling back to the Kilo Gateway; a qualified `provider/model` ID is also accepted. Prepared sessions without an initial prompt use the normal model defaults. Use `versions: true` only when the tasks are alternate versions of the same work to compare; otherwise, multiple tasks start as independent sessions. +Starting another session in an existing worktree reuses its files and branch without rerunning setup. Coordinate edits between sessions sharing that checkout. + +Each request can start up to 20 tasks. Sessions started with a prompt inherit the current turn's model and reasoning effort unless you explicitly request a different selection. Sessions created without a prompt use the normal defaults. Ask for independent sessions for separate tasks, or alternate versions of the same task to compare approaches. -The companion `agent_manager_models` tool searches models and their supported reasoning variants on demand. Results are grouped by model name (with the offering providers listed for reference) and limited to 20 per call, so the full catalog is never added to the conversation context. +The agent can use `agent_manager_models` to look up available models, providers, and reasoning effort before applying your requested override. The same tool also manages existing sessions. Use `action: "list"` first to get the current sections, worktrees, local sessions, and exact IDs. It can then send a prompt to one managed session, stop a managed session, move a worktree into a section, or answer a pending question. Moving accepts a section ID from the overview, or `null` to ungroup the worktree. Moving a session moves its whole worktree, including multi-version siblings. Local sessions cannot be assigned to a section. Stopping aborts the session's active work and removes it from the panel, just like closing the session tab. Prompts to busy or retrying sessions enter the same queue as follow-up messages sent from chat. The tool returns when the prompt is accepted, without waiting for it to run or finish. Sessions with pending questions or permission requests still refuse prompts. Answer a pending question with `action: "answer"`, or resolve the permission request in Agent Manager, before prompting again. A prompt targets one session; the tool does not broadcast to every session. -The tool uses the `agent_manager` permission. Approval prompts are scoped to the requested capability, so approving `worktree` does not automatically approve `local`, an overview, or a targeted prompt. Prompting an existing managed session requires an explicit `prompt` approval the first time, even if Agent Manager session creation was previously approved broadly. Stopping a session likewise requires an explicit `stop` approval, and moving a worktree requires an explicit `move` approval. +The tool uses the `agent_manager` permission. Approval prompts are scoped to the requested capability, so approving `worktree` does not automatically approve `local`, an overview, or a targeted prompt. Prompting an existing managed session requires an explicit `prompt` approval the first time, even if Agent Manager session creation was previously approved broadly. Stopping, moving, and answering a question each require their own approval. An answer must use the question's exact option labels; approving a plan can start implementation. ### Typical tool sequence @@ -399,6 +421,8 @@ Use `Cmd+T` / `Ctrl+T` in the panel, or `mode: "local"` with a selected `worktre - **Sessions in different worktrees:** Use targeted prompts plus commits, diffs, or pull requests to pass changes between isolated checkouts. Files are not shared automatically. - **Task descendants:** A `task` child belongs to the session that launched it. Its Kilo Swarm board is scoped to that session tree, not to every Agent Manager session in the project. +Messages sent by another agent show **Sent by Kilo from another session**. Select the source link to open that session. If it is closed, the message shows **Session not open** instead. + ## Sections Sections let you group worktrees into collapsible, color-coded folders in the sidebar. Use them to organize your workflow however you like — by status ("Review Pending", "In Progress"), by project area ("Frontend", "Backend"), priority, or any other scheme that fits. @@ -463,6 +487,8 @@ The worktree creation base and the diff comparison base are separate. The Branch - Markdown files include an eye/code toggle in the file header to switch between rendered Markdown and the raw diff - **Drag file headers into chat** — drag a file header from the diff panel into the chat input to insert an `@file` mention, giving the agent context about specific changed files +Files marked `linguist-generated` in the repository's `.gitattributes` start collapsed. Expand them to review their changes; files explicitly marked as not generated stay expanded. + ### Sending review comments Add comments in the diff panel or in the rendered view of a Markdown document. Click **Send all to chat** to send the collected comments to chat. If an Agent Manager terminal is active, the comments are sent to that terminal instead. Press `Cmd+Enter` (macOS) or `Ctrl+Enter` (Windows/Linux) to use the same action from the review panel. @@ -501,6 +527,16 @@ In a rendered Markdown preview, use the comment control in the line gutter to ad The project and worktree context owns document tabs, loaded content, and comments. The session ID attached to an opened file selects the session's worktree for reading and native-editor navigation. Sessions that share one worktree also share its document inspector state; switching project or worktree changes the visible context without mixing tabs or comments across worktrees. +## Browser previews + +Enable **Browser Automation** in **Settings > Experimental** to show the Browser panel. It is off by default. Open the panel with the **Browser** button and enter your local application's URL. Each session has its own browser, with developer tools and console diagnostics beside the preview. + +Use **Select element** to attach an element reference to your next message. It includes enough page context for Kilo to identify the element, plus a source file and line when the page provides a verifiable location. Add your instruction before sending it. + +The agent can open the application with `browser_open`. Its automation browser accepts only HTTP URLs on `localhost` or `127.0.0.1` and blocks other origins, including external assets and APIs on separate ports. Serve the resources the agent needs from the same loopback origin. These restrictions apply to the automation browser, not the independent visible preview iframe. + +Google Chrome must be installed for the default runtime. To use an already-installed compatible Playwright Chromium browser instead, turn off **Use System Chrome** under **Settings > Web Tools**. + ## Terminals Each session has a dedicated terminal rooted in the session's worktree directory. Press `Cmd+/` (macOS) / `Ctrl+/` (Windows/Linux) to focus the terminal for the active session. If the embedded terminal is already visible but the prompt has focus, the same shortcut focuses the terminal without hiding it. Press it again while the terminal has focus to hide the panel. @@ -511,8 +547,8 @@ When you use `@terminal` in an Agent Manager prompt, Kilo captures the focused t The toolbar's terminal button is a split button: click it to open a terminal, or use its dropdown to choose where terminals open: -- **VS Code terminal** (default) — opens or focuses the VS Code integrated terminal at the bottom of the window -- **Agent Manager panel** — opens an embedded terminal in the side panel that also hosts the diff view, so the shell stays inside the Agent Manager layout +- **Agent Manager panel** (default) - opens an embedded terminal in the side panel that also hosts the diff view +- **VS Code terminal** - opens or focuses the VS Code integrated terminal at the bottom of the window The dropdown choice is remembered per panel and becomes the default for new panels. You can also set the default directly with the `kilo-code.new.agentManager.terminalButtonDestination` setting (`vscode` or `agentManager`). The `Cmd+/` (macOS) / `Ctrl+/` (Windows/Linux) shortcut follows the same destination. @@ -657,6 +693,10 @@ Two extra variables are injected into the script's environment: The terminal destination dropdown in the Agent Manager toolbar also controls where the script runs. **Agent Manager panel** uses the named side terminal, while **VS Code terminal** runs it as a task in the integrated terminal. The integrated terminal option is kept for comparison and will be removed in a future release. +## Keep Awake + +Use the coffee-cup button in the project list header, or **Kilo Code: Toggle Keep Awake** in the Command Palette, to prevent system sleep while agents work. It starts off in each window and requires a trusted local workspace. See [Keep Awake](/docs/getting-started/settings/keep-awake) for platform limits, reset behavior, and device-policy considerations. + ## Session State and Persistence Agent Manager state is persisted in `.kilo/agent-manager.json`. It stores worktrees, sections, session tabs, ordering, collapsed state, diff preferences, and cached PR metadata. Git branches and worktree directories remain on disk separately. diff --git a/packages/kilo-docs/pages/automate/code-reviews/github.md b/packages/kilo-docs/pages/automate/code-reviews/github.md index 753a25ffbeb3..afdd9ad3d22f 100644 --- a/packages/kilo-docs/pages/automate/code-reviews/github.md +++ b/packages/kilo-docs/pages/automate/code-reviews/github.md @@ -79,14 +79,19 @@ When a review triggers: The repository list is synced from GitHub and can be refreshed from the configuration page. +### Per-repository overrides + +Turn automated reviews on or off for an individual repository to override the installation default. For example, disable a noisy repository or enable one repository while the default is off. Repositories with reviews disabled do not start automated reviews for incoming pull requests. + ## Troubleshooting ### Reviews are not triggering 1. Verify the GitHub App is installed and has access to the repository -2. Check that the Review Agent is **enabled** in the Code Reviews configuration +2. Check that reviews are enabled by the installation default or a per-repository override 3. Ensure the repository is in the allowed list (if using "Selected repositories" mode) -4. Confirm the PR is not a draft +4. Check that a per-repository override has not disabled reviews +5. Confirm the PR is not a draft ### Reviews are failing diff --git a/packages/kilo-docs/pages/automate/code-reviews/overview.md b/packages/kilo-docs/pages/automate/code-reviews/overview.md index e4d35150fe8e..75a936799c85 100644 --- a/packages/kilo-docs/pages/automate/code-reviews/overview.md +++ b/packages/kilo-docs/pages/automate/code-reviews/overview.md @@ -124,6 +124,7 @@ Use `/review` for all local code reviews: - **`/review`** — Review uncommitted changes (staged, unstaged, and untracked) when run without arguments - **`/review uncommitted [guidance]`** — Review uncommitted changes with optional guidance - **`/review branch [base] [guidance]`** — Review your current branch vs. its detected or specified base, with optional guidance +- **`/review worktree [guidance]`** - In an Agent Manager worktree session, review committed, staged, unstaged, and untracked changes against the recorded parent branch - **`/review `** — Review a specific commit - **`/review `** — Review a pull request diff --git a/packages/kilo-docs/pages/automate/integrations.md b/packages/kilo-docs/pages/automate/integrations.md index fb2146f314ef..036947ef2008 100644 --- a/packages/kilo-docs/pages/automate/integrations.md +++ b/packages/kilo-docs/pages/automate/integrations.md @@ -30,7 +30,7 @@ Before connecting: - For GitHub: You need permission to install GitHub Apps for the repositories you want Kilo to access. - For GitLab: You need **Maintainer** role (or higher) on the projects you want to connect. - For DoltHub: You need a DoltHub account to authorize the OAuth connection. -- (Optional) If you're connecting an organization, you must be an admin or have app installation permissions. +- For a Kilo organization, adding a GitHub integration requires the **Owner** or **Admin** role, in addition to permission to install the app on GitHub. --- @@ -161,9 +161,9 @@ Once your integrations are connected, the following features are enabled in Kilo - Use DoltHub alongside GitHub or GitLab when a workflow also needs repository access - Authorize Gas Town Wasteland to fork commons databases, push claims and evidence, and manage DoltHub PRs. Wasteland also supports an advanced API token option when OAuth is not available. -### Upcoming: +### Bitbucket -- **Bitbucket Integration** +Organization accounts can connect Bitbucket Cloud and select its repositories when starting [Cloud Agent sessions](/docs/code-with-ai/platforms/cloud-agent) on the web or in the mobile app. Connect Bitbucket from your organization's **Integrations** page. Bitbucket is not available for personal accounts. --- @@ -171,6 +171,8 @@ Once your integrations are connected, the following features are enabled in Kilo ### GitHub +An organization Owner or Admin can connect multiple GitHub organizations. Click **Add organization** and repeat the installation flow. Each installation has its own connection status, repository scope, selected repository list, and management actions. Personal accounts use a single GitHub connection. + From the **Integrations** page, click "Manage on GitHub" to: - View the GitHub account you connected diff --git a/packages/kilo-docs/pages/automate/tools/index.md b/packages/kilo-docs/pages/automate/tools/index.md index 78d2a43f6866..d7d9f1825e8e 100644 --- a/packages/kilo-docs/pages/automate/tools/index.md +++ b/packages/kilo-docs/pages/automate/tools/index.md @@ -19,7 +19,7 @@ Tools are organized into logical groups based on their functionality: | **Edit Group** | File system modifications | `edit`, `write`, `apply_patch` | Code changes and file manipulation | | **Execute Group** | Shell command execution | `bash` | Running scripts, building projects | | **Web Group** | Fetch and search web content | `webfetch`, `websearch` | Research, documentation lookup | -| **Browser Group** | Web browser automation | `kilo-playwright_*` (via built-in Playwright MCP) | Browser testing and interaction | +| **Browser Preview** | Agent Manager's built-in browser preview | `browser_open` | Previewing and checking locally running apps | | **MCP Group** | External tool integration | MCP server tools (namespaced as `{server}_{tool}`) | Specialized functionality via MCP | | **Workflow Group** | Sub-agents and task management | `question`, `task`, `todowrite`, `todoread`, `plan`, `skill`, `agent_manager`, `board_post`, `board_read` | Context switching and task organization | @@ -94,15 +94,11 @@ Set the `KILO_WEBSEARCH_PROVIDER` environment variable to force a provider: ### Browser Tools -The VS Code extension has a built-in browser automation tool powered by [Playwright MCP](https://www.npmjs.com/package/@playwright/mcp). Enable it in Settings → Browser Automation. When enabled, it registers an MCP server named `kilo-playwright` and exposes tools such as: +The VS Code extension's experimental `browser_open` tool opens a local application in Agent Manager's Browser panel and returns a screenshot and diagnostics. Enable **Browser Automation** under **Settings > Experimental**. It requires installed Chrome or compatible Playwright Chromium. -- `kilo-playwright_browser_navigate` - Navigate to a URL -- `kilo-playwright_browser_click` - Click an element -- `kilo-playwright_browser_type` - Type text into an element -- `kilo-playwright_browser_screenshot` - Capture a screenshot -- `kilo-playwright_browser_snapshot` - Capture an accessibility snapshot +The `browser_open` automation browser accepts HTTP URLs on `localhost` or `127.0.0.1` only, and blocks resources from other origins. See [Browser previews](/docs/automate/agent-manager#browser-previews) for setup and element feedback. -These follow the same permission model as all MCP tools (see below). +This restriction is specific to `browser_open`, not Kilo's web access in general. Use `websearch` and `webfetch` to find and read public web pages. Browser tools from a separately configured MCP server can provide interactive web browsing according to that server's capabilities and permissions. ### MCP Tools @@ -118,6 +114,7 @@ These tools help manage the conversation and task flow: - `todoread` - Reads the current session TODO list - `plan` - Enters structured planning mode - `skill` - Invokes a reusable skill (Markdown instruction module) +- `open_plan` - Opens a saved plan for review in the VS Code extension - `agent_manager` - Starts Agent Manager local or worktree sessions in VS Code - `board_post` / `board_read` - Exchange messages on the experimental Kilo Swarm board diff --git a/packages/kilo-docs/pages/code-with-ai/agents/context-mentions.md b/packages/kilo-docs/pages/code-with-ai/agents/context-mentions.md index 2a88ed2de4cf..bbc81c70704a 100644 --- a/packages/kilo-docs/pages/code-with-ai/agents/context-mentions.md +++ b/packages/kilo-docs/pages/code-with-ai/agents/context-mentions.md @@ -26,6 +26,8 @@ Type `@` in the chat input to get autocomplete suggestions. You can mention: | **Terminal** | In the sidebar, include the active VS Code terminal output. In Agent Manager, include the focused terminal for the selected session or worktree, including embedded **Run** and **Setup** tabs. | `@terminal` | | **Git Changes** | Attach uncommitted working-tree diffs and new files | `@git-changes` | | **Past chats** | Add a previous session's chat history as context | `@` → **Past chats** → pick a session | +| **Worktrees** | In Agent Manager, reference another worktree's location and session metadata | `@` → **Worktrees** → pick a worktree | +| **Model** | Reference a model for a subagent or Agent Manager session | `@` → **Model** → pick a model | Selecting a suggestion inserts the mention and highlights it in the input. File contents, terminal output, and git changes are attached as context when you send the message. @@ -37,6 +39,12 @@ Choosing **Past chats** from the `@` menu opens a searchable picker of sessions Very long chat histories are truncated, keeping the beginning and end, so a single mention cannot fill the context window. +### Referencing worktrees and models + +In Agent Manager, choose **Worktrees** to search other worktrees by name, branch, or session title. A worktree mention includes its path, branch, base branch, and session metadata. It does not attach files, diffs, or chat history, and does not change your working directory. + +Choose **Model** to insert an inline `@provider/model` reference. Use it to request a particular model for a subagent or Agent Manager session. It does not attach a file or change the current session's model. Subagent overrides require [Task Subagent Model Selection](/docs/getting-started/settings#task-subagent-model-selection). + ### Drag and Drop You can also add file mentions by dragging and dropping: diff --git a/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md b/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md index d511415d6948..23e863f6b228 100644 --- a/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md +++ b/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md @@ -11,10 +11,10 @@ Instead of maintaining a static list that's perpetually behind, we built somethi ## Model Routing and Configuration -Kilo's IDE Extension and CLI configurations have four separate, independently configurable model slots: +You can configure models separately for different tasks in the VS Code extension and kilo CLI: - **Main model** — the primary model your agent uses for coding tasks, chat, and reasoning. This is what you pick with the model selector, `/models`, or the `model` key in `kilo.jsonc`. See [How to Select and Switch Models](#how-to-select-and-switch-models) for the full precedence order and per-agent config. - - The main model is also used for context compaction/summarization and todo-list generation. + - The main model is also used for todo-list generation, and for context compaction when no compaction model is set. - **Small model** — a lightweight model used for session title generation, commit message generation, and prompt enhancement. Configured with the `small_model` key in `kilo.jsonc`, or the **Small Model** field on the **Settings → Models** tab. - If left unset, Kilo resolves it according to the following logic: 1. Find a small/cheap variant on your current provider (e.g. Haiku on Anthropic, Flash on Gemini). @@ -23,6 +23,7 @@ Kilo's IDE Extension and CLI configurations have four separate, independently co - **Subagent model** — the default model for subagents launched by the `task` tool. Configured with the `subagent_model` key in `kilo.jsonc`, or the **Subagent Model** field on the **Settings → Models** tab. - If left unset, inherits whichever model the parent agent session is currently using. - **Autocomplete model** — the model used for inline code completions as you type. See [Autocomplete: Provider and Model](/docs/code-with-ai/features/autocomplete#provider-and-model) for how to configure it. +- **Compaction model** - the model used to summarize context. Set `agent.compaction.model` in `kilo.jsonc`, or choose **Compaction model** under **Settings → Models**. If unset, compaction uses the current session's model. See [Context Condensing](/docs/customize/context/context-condensing#use-a-different-model-for-compaction). ### Configuring Local Usage @@ -153,6 +154,12 @@ The Settings UI writes the same `agent..model` entry, so either method pro For details on configuring subagent models, see [Custom Subagents](/docs/customize/custom-subagents). +### Per-task model selection (experimental) + +Enable [Task Subagent Model Selection](/docs/getting-started/settings#task-subagent-model-selection) to request a different model, provider, or reasoning effort for an individual subagent task. It is off by default. Overrides require your explicit request; the agent does not choose a different model on its own based on task complexity, cost, or latency. + +Invalid selections fail instead of silently falling back. Resumed tasks keep their last model and reasoning effort unless you request an override. + ## Selecting a Model or Agent via a Link (VS Code) The VS Code extension supports a `vscode://` protocol handler that lets you open VS Code and automatically select a model, an agent, or both — no manual picker interaction required. This is useful for sharing model recommendations, launching a specific model tier from a web page, or switching quickly to a preferred agent. diff --git a/packages/kilo-docs/pages/code-with-ai/agents/orchestrator-mode.md b/packages/kilo-docs/pages/code-with-ai/agents/orchestrator-mode.md index 1f1bead1d4d6..edb56792fb23 100644 --- a/packages/kilo-docs/pages/code-with-ai/agents/orchestrator-mode.md +++ b/packages/kilo-docs/pages/code-with-ai/agents/orchestrator-mode.md @@ -25,7 +25,9 @@ Now, **subagent support is built into agents that have full tool access** (Code, 1. The agent analyzes a complex task and decides a subtask would benefit from isolation. 2. It launches a subagent session using the `task` tool (e.g., `general` for autonomous work, `explore` for codebase research). -3. The subagent runs in its own isolated context — separate conversation history, no shared state. -4. When done, the subagent returns a summary to the parent agent, which continues its work. +3. The subagent has a separate conversation history but shares the parent's project directory or worktree. It does not isolate file edits. +4. A foreground task returns its result before the parent continues. A background task lets the parent continue immediately and delivers its result later. Agents can launch multiple subagent sessions concurrently for parallel work. + +[Kilo Swarm](/docs/getting-started/settings#kilo-swarm) lets a main session and its task descendants exchange findings on a shared board. It works with your current agent and does not require the deprecated Orchestrator mode. Swarm is on by default; turn it off in **Settings > Agent Behaviour** or set `experimental.shared_agent_board` to `false` in `kilo.jsonc`. For separate branches and checkouts, use [Agent Manager worktree sessions](/docs/automate/agent-manager#orchestration-model) instead. diff --git a/packages/kilo-docs/pages/code-with-ai/agents/using-agents.md b/packages/kilo-docs/pages/code-with-ai/agents/using-agents.md index 7f6ddf8cfc77..649640185f9f 100644 --- a/packages/kilo-docs/pages/code-with-ai/agents/using-agents.md +++ b/packages/kilo-docs/pages/code-with-ai/agents/using-agents.md @@ -68,7 +68,7 @@ There are several ways to switch agents: | **Description** | An experienced technical leader and planner who helps design systems and create implementation plans | | **Tool Access** | Read-only tools plus restricted file editing (plan files in `.kilo/plans/` only) | | **Ideal For** | System design, high-level planning, and architecture discussions | -| **Special Features** | Similar to the legacy extension's "Architect" mode, with a planning-focused approach | +| **Special Features** | Similar to the legacy extension's "Architect" mode, with a planning-focused approach. In VS Code, the saved plan opens in the editor when ready for review. | ### debug diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/cli-reference.md b/packages/kilo-docs/pages/code-with-ai/platforms/cli-reference.md index 56782a8c6abd..cf1fd640ea09 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/cli-reference.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/cli-reference.md @@ -1028,30 +1028,36 @@ Options: start a Cloud Agent task Options: - --help Show help [boolean] - --version Show version number [boolean] - --prompt prompt for the Cloud Agent [string] [required] - --repo repository shorthand or URL [string] - --repo-type repository provider type [string] [choices: "github", "gitlab", "git"] - --branch repository branch [string] - --model Cloud Agent model [string] - --mode Cloud Agent mode [string] - --org-id Kilo organization ID [string] - --stream connect to the WebSocket stream and print events as JSONL [boolean] + --help Show help [boolean] + --version Show version number [boolean] + --prompt prompt for the Cloud Agent [string] + --prompt-stdin read the prompt from standard input [boolean] [default: false] + --repo repository shorthand or URL [string] + --repo-type repository provider type [string] [choices: "github", "gitlab", "git"] + --branch repository branch [string] + --model Cloud Agent model [string] + --mode Cloud Agent mode [string] + --org-id Kilo organization ID [string] + --stream connect to the WebSocket stream and print events as JSONL [boolean] ``` +Provide exactly one of `--prompt` or `--prompt-stdin`. For example, read a multiline prompt from a file with `kilo cloud start --repo owner/repo --prompt-stdin < prompt.txt`. + ### kilo cloud send ``` send a follow-up prompt to a Cloud Agent task Options: - --help Show help [boolean] - --version Show version number [boolean] - --session-id Cloud Agent session ID [string] [required] - --prompt follow-up prompt for the Cloud Agent [string] [required] + --help Show help [boolean] + --version Show version number [boolean] + --prompt prompt for the Cloud Agent [string] + --prompt-stdin read the prompt from standard input [boolean] [default: false] + --session-id Cloud Agent session ID [string] [required] ``` +Provide exactly one of `--prompt` or `--prompt-stdin`, as with `kilo cloud start`. + ### kilo cloud status ``` diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/cli.md b/packages/kilo-docs/pages/code-with-ai/platforms/cli.md index a12731d20881..9d53a11f2cf2 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/cli.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/cli.md @@ -131,6 +131,7 @@ The `kilo console` command and its browser interface are deprecated and will be | Command | Aliases | Description | |---|---|---| | `/status` | - | View status | +| `/about` | - | Show version, runtime, configuration paths, and provider details; press `c` to copy diagnostics | | `/themes` | - | Switch theme | | `/help` | - | Show help | | `/reload` | - | Reload every instance of the project from disk (config, skills, agents, and commands) | @@ -246,7 +247,7 @@ There is no slash command or command-palette toggle for notifications or sounds. The CLI's interactive mode supports slash commands for common operations. The main commands are documented above in the [Interactive Slash Commands](#interactive-slash-commands) section. -Use `/diff` to review working-tree changes. From the diff viewer, switch the source to the current branch compared with the main branch or to changes from the last assistant turn. Use `/move` to move the current session to another project directory. +Use `/diff` to review working-tree changes. From the diff viewer, switch the source to the current branch compared with the main branch, changes from the last assistant turn, or the last commit (`HEAD` vs `HEAD~1`). Use `/move` to move the current session to another project directory. The `diff_open` and `session_move` TUI keybindings run the same actions and are unbound by default. Set them under `keybinds` in `tui.jsonc`: @@ -603,7 +604,9 @@ This instructs the AI to proceed without user input. - `0`: Success (task completed) - `124`: Timeout (task exceeded time limit) -- `1`: Error (initialization or execution failure) +- `1`: Error (initialization, execution, or request failure) + +A run that finishes without an assistant message also exits `1`, reporting `run ended without an assistant message; the model returned no output` on stderr or as a final `error` record with `--format json`. If the prompt request fails, Kilo reports that error instead. Without `--auto`, a non-interactive run cannot prompt for approval and auto-rejects any permission request it receives. If a run auto-rejected at least one request, it exits `1` with a stderr diagnostic naming the cause, since the task likely did not complete. Pass `--auto` for autonomous use. @@ -682,6 +685,8 @@ Add to `~/.config/kilo/config.json`: Once enabled, start a CLI session and open [Cloud Agents](https://app.kilo.ai/cloud). Your local session appears in the dashboard. See [Cloud Agent Remote Connections](/docs/code-with-ai/platforms/cloud-agent#remote-connections) for details. +A connected client can start a session in a child folder of the CLI's launch directory, or continue an existing cloud session locally. Requested folders must remain within the launch directory; absolute paths and paths that escape it are rejected. + ### Requirements - Connection to Kilo Gateway diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/cloud-agent.md b/packages/kilo-docs/pages/code-with-ai/platforms/cloud-agent.md index ddbe85f30798..3ca832833b6b 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/cloud-agent.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/cloud-agent.md @@ -5,12 +5,12 @@ description: "Using Kilo Code in the browser" # {% $markdoc.frontmatter.title %} -Cloud Agents let you run Kilo Code in the cloud from any device, without relying on your local machine. They provide a remote development environment that can read and modify your GitHub and GitLab repositories, run commands, and auto-commit changes as work progresses. +Cloud Agents let you run Kilo Code in the cloud from any device, without relying on your local machine. They provide a remote development environment that can read and modify your GitHub, GitLab, or Bitbucket repositories, run commands, and optionally commit and push changes as work progresses. Bitbucket repositories are available for organizations only. ## What Cloud Agents Enable - Run Kilo Code remotely from a browser -- Auto-create branches and push work continuously +- Create branches and optionally commit and push changes automatically - Use env vars + startup commands to shape the workspace - Work from anywhere while keeping your repo in sync @@ -18,8 +18,8 @@ Cloud Agents let you run Kilo Code in the cloud from any device, without relying Before using Cloud Agents: -- **GitHub or GitLab Integration must be configured** - Connect your account via the [Integrations tab](https://app.kilo.ai/integrations) so that Cloud Agents can access your repositories. +- **A GitHub, GitLab, or Bitbucket integration must be configured** + Connect your account via the [Integrations tab](https://app.kilo.ai/integrations) so that Cloud Agents can access your repositories. For Bitbucket, use your organization's **Integrations** page; its repositories are available in both web and mobile Cloud Agent sessions. ## Cost @@ -35,14 +35,20 @@ Usage is measured in whole seconds, with no rounding up to a longer billing inte See [Kilo Code pricing](https://kilo.ai/pricing) for current rates and pricing for other cloud products. +### Session cost display + +The session header separates **Token Usage** (model inference spend) from **Compute** (the cloud environment's estimated hourly rate). Several sessions can share that environment, so the rate is not a per-session spending total. **Not currently charged** means compute charges are not being applied. + +If a compute billing check fails, follow the recovery action next to the composer. Available actions depend on whether you or your organization pays and your role in that organization. Your prompt is preserved so you can retry after resolving the issue. + ## How to Use -1. **Connect your GitHub or GitLab account** in the [Integrations](https://app.kilo.ai/integrations) tab of your personal or organization dashboard. +1. **Connect your GitHub, GitLab, or Bitbucket account** in the [Integrations](https://app.kilo.ai/integrations) tab of your personal or organization dashboard. Bitbucket requires an organization. 2. **Select a repository** to use as your workspace. 3. **Add environment variables** (secrets supported) and set optional startup commands. 4. **Start chatting with Kilo Code.** -Your work is always pushed to GitHub, ensuring nothing is lost. +Automatic commit and push depends on your session settings. If it is disabled, review, commit, and push the changes you want to keep. ## Starting Tasks from the CLI @@ -54,18 +60,26 @@ kilo cloud start --prompt "Fix the flaky login test" --repo Kilo-Org/kilocode `kilo cloud` can start tasks, send follow-up prompts, and check task status and results. Repository, branch, model, mode, and organization are inferred from your local checkout and CLI defaults unless you pass the matching flags. Add `--stream` to `kilo cloud start` to print task events as JSONL until the task completes. See the [CLI reference](/docs/code-with-ai/platforms/cli-reference#kilo-cloud) for all commands and options. +`kilo cloud start` and `kilo cloud send` require exactly one prompt source: `--prompt` or `--prompt-stdin`. Use `--prompt-stdin` to read a file or another command's output from standard input: + +```bash +kilo cloud start --prompt-stdin --repo Kilo-Org/kilocode < task.md +``` + +Standard-input prompts must be valid UTF-8 and contain no more than 100,000 characters. + ## How Cloud Agents Work - Each user receives an **isolated Linux container** with common dev tools preinstalled (Node.js, git, gh CLI, glab CLI, etc.). - Python is not included in the base image, but `apt` is available so you can install it or other packages as needed. -- All Cloud Agent chats share a **single container instance**, while each session gets its own workspace directory. -- When a session begins: +- All Cloud Agent chats share a **single container instance**. Each worktree has its own checkout directory, and chats in the same worktree share it. +- When a new worktree is created: 1. Your repo is cloned 2. A unique branch is created 3. Your startup commands run 4. Env vars are injected -- After every message, the agent: +- When automatic commit and push is enabled, after each message the agent: - Looks for file changes - Commits them - Pushes to the session’s branch @@ -75,6 +89,28 @@ kilo cloud start --prompt "Fix the flaky login test" --repo Kilo-Org/kilocode - Expect slightly longer setup after idle periods - Inactive cloud agent sessions are deleted after **7 days** during the beta, expired sessions are still accessible via the CLI +## Worktrees and chats + +A worktree is a checkout of your repository on its own branch. It can host multiple chats with separate conversations, but edits from one chat are visible to the others. Use separate worktrees for tasks that need separate checkouts. + +Each chat opens as its own tab in the worktree group. You can rename tabs, close and reopen them, or delete a whole worktree and all of its chats. + +### Workspace folders + +Group worktrees into folders in the sidebar. Folders are private to you, not shared with your organization. + +- Create a folder with a name and an optional color, then drag worktrees onto it. +- Use the folder menu to rename it, change its color, move it up or down, or delete it. Deleting a folder returns its worktrees to **Ungrouped** without deleting them. +- Click a folder's header to collapse or expand it. + +## Reviewing changes + +Select **Changes** in the chat header to review the worktree's saved change summary, including file status and lines added and removed. The panel shows the comparison's base branch and when the summary was saved. Refresh the panel to load the latest saved summary. Large summaries can be partial, with some files or line counts omitted. + +Select a file to open its saved diff and, when available, full contents in a read-only tab. Reloading reads the latest saved revision without starting the workspace. + +These are saved snapshots, not a live view of the checkout. Diffs or contents may be unavailable for binary, unsupported, or large files, or when there are too many changes to save in full. + ## Agent Environment Profiles Agent environment profiles are reusable bundles of environment settings for cloud-agent sessions. A profile can include: @@ -185,6 +221,8 @@ When creating a trigger, you choose an **activation mode** that cannot be change Triggers utilize [agent environment profiles](#agent-environment-profiles) to configure the execution environment for triggered sessions. The agent resolves the profile at runtime, so profile updates apply automatically to future executions. Profiles referenced by triggers cannot be deleted until those triggers are updated or removed. +For models that support it, webhook and scheduled triggers can set a **reasoning effort** next to the model selector. **Default** leaves the model's behavior unchanged; a specific effort applies to each session the trigger starts. + Triggers do not support manual env var or setup command overrides at this time. ### Scheduled Triggers @@ -193,6 +231,12 @@ Scheduled triggers fire on a recurring schedule using cron expressions. You can The minimum schedule interval is 10 minutes. Scheduled triggers use `{{scheduledTime}}` and `{{timestamp}}` as prompt template variables (webhook-specific variables like `{{body}}` are not available since there is no inbound HTTP request). +### Invoking a scheduled trigger on demand + +To run an active scheduled trigger immediately, select the **Play** action on its row and confirm. When editing one, use **Save and invoke now** to save your changes before starting the run. Paused triggers cannot be invoked, and webhook triggers have no invoke action. + +Manual invocations use the same queue and in-flight limits as scheduled runs. Invoking alone does not change the cron schedule, timezone, or next scheduled occurrence; any schedule edits saved with **Save and invoke now** still take effect. + ### Trigger Limits and Guidance Triggers are designed for low-volume invocations from trusted sources and are best suited for short-lived tasks. diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/jetbrains.md b/packages/kilo-docs/pages/code-with-ai/platforms/jetbrains.md index 02d489552e82..7a7d93c918bd 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/jetbrains.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/jetbrains.md @@ -11,18 +11,75 @@ description: "Using Kilo Code in JetBrains IDEs" ## Settings -Open **Settings → Tools → Kilo Code** to configure the plugin. The JetBrains plugin reads and writes the same shared `kilo.jsonc` config files as the CLI and the VS Code extension, so changes apply across clients. See [Settings](/docs/getting-started/settings) for config file locations and precedence. +Open **Settings → Tools → Kilo Code** to configure the plugin. Shared agent settings use the same `kilo.jsonc` files as the CLI and VS Code extension; IDE-specific options such as GitHub integration and worktree indexing stay in the plugin. See [Settings](/docs/getting-started/settings) for config file locations and precedence. - **Auto-Approve** — set per-tool permission levels (Allow / Ask / Deny) and manage granular command and path exceptions without editing config by hand. Permission prompts offer one-time approvals alongside saved allow/reject rules. See [Auto-Approving Actions](/docs/getting-started/settings/auto-approving-actions) for the shared permission model. - **Context** — toggle auto-compaction, set the auto-compaction limit (the percentage of the model window that triggers compaction), enable pruning of old tool outputs, and manage file watcher ignore patterns. See [Context Condensing](/docs/customize/context/context-condensing) and [.kilocodeignore](/docs/customize/context/kilocodeignore) for what these settings control. - **Agent Behavior → Skills** — inspect loaded skills, add extra skill sources (local paths or remote URLs), edit or remove custom skills, and open skill files in the editor. See [Skills](/docs/customize/skills) for the skill format and discovery rules. +- **Integrations** - enable or disable the GitHub integration for pull request badges and imports. It requires the GitHub CLI (`gh`) to be installed and authenticated. +- **Advanced → Index agent worktrees** - include `.kilo/worktrees` in the containing project's index. Worktrees are excluded by default to avoid duplicate search results. Files opened from an excluded worktree in the main IDE window lack code resolution and inspections; open the worktree as its own project for full indexing. + +## Chat and worktrees + +Use **Chat** for the current workspace and **Agents** to manage parallel tasks in isolated git worktrees. **+ Session** starts a conversation; **+ Worktree** opens the worktree creation dialog. + +- **New Worktree** creates a new branch, imports a GitHub pull request with **From PR**, or uses an available local branch with **From Branch**. +- **Move to Worktree** moves the conversation and uncommitted changes into a new worktree while the session is idle. This action is also available from the main checkout's session list. +- Open a worktree to see its sessions in an editor tab. Its session list is scoped to that worktree; use the list toggle to hide or show it and drag worktree rows to reorder them. +- Worktree rows show session activity, pull request checks and reviews, unresolved review conversations, merge conflicts, and active build/run processes. Use the row menu to copy the branch name, directory, or pull request reference. + +### Worktree setup scripts + +Add a setup script to install dependencies or prepare configuration in new worktrees. Kilo starts it automatically in a terminal, but **does not wait for it to finish before starting the session**. Wait for setup to complete before asking the agent to use those dependencies or generated files. + +| Platform | Filename (checked in order) | +|---|---| +| macOS / Linux | `.kilo/setup-script`, `.kilo/setup-script.sh` | +| Windows | `.kilo/setup-script.ps1`, `.kilo/setup-script.cmd`, `.kilo/setup-script.bat` | + +The terminal runs in the worktree directory with `WORKTREE_PATH` (the worktree directory) and `REPO_PATH` (the repository root) available as environment variables. Use the worktree row menu to create or open the script, or choose **Run Worktree Setup** to run it again. + +### Running code in a worktree + +Open **Build/Run** in the worktree editor to choose a supported IDE run configuration. Eligible Application, Spring Boot, and Kotlin/Groovy application configurations can run through the project's build system using the worktree's code. Support depends on the configuration and build-system integration; the popup identifies the build system used. A plain-application fallback may omit framework settings, which Kilo reports in a notification. + +Use **Show Output** to view a running process's console, **Stop** to stop it, or **Kill** if it remains running. **Build** and **Rebuild** are available for supported build systems. Removing a worktree stops its running processes. For unsupported configurations or full IDE run and debug support, choose **Open in New Frame**. + +### Forking a session + +Use **Fork Session** in a worktree session's row menu, right-click menu, or prompt bar's more menu to try another approach without losing the original conversation. To branch from an earlier message, use that user message's hover toolbar. The copied conversation opens as a new session next to its source; forking does not create a separate worktree. + +## Diagrams in chat + +Ask Kilo for a Mermaid diagram to visualize a workflow, architecture, data relationship, or timeline. Chat renders `mermaid` and `mmd` code blocks inline, with source shown while streaming or if rendering fails. + +Click a diagram to open a zoomable viewer, or use its toolbar to open an editor tab with **Diagram** and read-only **Source** views. Copying a rendered diagram copies a PNG; copying while it is still streaming or after a render error copies the source instead. ## Reviewing session changes - **Modified files per turn** — each assistant turn that changed files shows a **Modified** card with the affected files and their diff stats. Expand a file to see its diff inline, or open all of the turn's changes in the **Changed files** diff viewer. -- **Branch comparison** — when the workspace differs from the base branch, the session header shows a changes badge. Click it (**Compare with base branch**) to open a diff editor with a file tree and per-file navigation. +- **Branch comparison** - use the session header's **Compare with base branch** badge to open a diff editor with a file tree and per-file navigation. A separate uncommitted-changes badge compares local edits with the last commit. - **Stale diff refresh** — diff views detect when files change on disk and offer a **Refresh** action to reload them instead of showing outdated content. +Worktree rows separate committed changes against the base branch from uncommitted changes. Select the uncommitted-changes badge to compare with `HEAD`, or use **Compare to Base** in the session menu to review the branch's changes including uncommitted work. + +## Session controls + +Right-click in a session or open the prompt bar's more menu to compare changes, copy the session ID, or share the conversation. **Share Session** creates a public link; **Stop Sharing** revokes it. Sharing requires signing in to Kilo and must be allowed by your configuration. The right-click menu also includes **Stop Session**. Its **Auto-Approve** toggle applies across Kilo sessions in the IDE, not just the current conversation. + +If a turn fails, use **Retry** after resolving the problem or selecting a different model or agent. Retry uses the current selections. A turn you stop yourself is marked as stopped, not as a failure. + +### Keyboard shortcuts + +These shortcuts work while a Kilo session is active. They use `Ctrl` on macOS too, and can be changed in **Settings → Keymap**. + +| Shortcut | Action | +|---|---| +| `Ctrl+1` | Cycle modes | +| `Ctrl+2` | Cycle favorite models, or recommended models if you have no favorites | +| `Ctrl+3` | Cycle reasoning effort for the current model | +| `Ctrl+0` | Reset the model override | + ## Permission requests When the agent asks for several approvals at once, permission requests queue up instead of replacing each other. Resolve the current request to advance to the next one in the queue. diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/mobile.md b/packages/kilo-docs/pages/code-with-ai/platforms/mobile.md index abd5c02a1a87..cf6e1fbb9013 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/mobile.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/mobile.md @@ -25,16 +25,36 @@ 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. +## Finding sessions + +The **Agents** tab shows live sessions. Tap **See all** there to search past sessions, filter by platform or project, and change the sort order. **See all** on Home opens the live Agents list instead. + ## 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 +### App lock + +Turn on **Unlock with biometrics** in **Settings > Preferences** to protect the app on this device. It asks you to unlock at launch and after five minutes in the background, with your device passcode available as a fallback. Locking hides your sessions and open sheets without discarding drafts or navigation. + +### Links and images in chat + +Images embedded in model replies stay hidden until you tap **Load**. The app shows the image's host first and only loads HTTPS images. + +Web links in model replies ask you to confirm the destination host before opening. Choose **Trust this host** to skip future confirmations for that host. Review or revoke trusted hosts in **Preferences > Trusted hosts**; signing out clears the list. This confirmation applies to HTTP and HTTPS links, not other link types. + +## Language and region + +The app follows your device language, with English as the fallback. Choose another language from the login screen or **Preferences**. The picker supports search and shows languages in their own scripts. Dates, times, and currency formatting follow the selected language. Switching between left-to-right and right-to-left layouts restarts the app. + +## 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). +On Android, you can buy, restore, and change Kilo Pass tiers through Google Play. Tier changes take effect at the next renewal; your current tier and credits stay in place until then. Google manages cancellation and payment methods for passes purchased through Google Play. + {% imageGallery columns="3" width="220px" %} {% image src="/docs/img/mobile-apps/home.webp" alt="Kilo Code mobile home screen showing active agent sessions" caption="Start coding tasks and resume active sessions from the mobile home screen." /%} @@ -52,9 +72,17 @@ For Kilo Pass pricing, billing, and account management details, use the [Kilo Pa The new-session screen includes a **Run on** picker that chooses where your session runs: - **Cloud Agent** — the managed cloud environment (the default). -- **A connected CLI instance** — a `kilo remote` CLI running on your own machine. The picker lists the instances currently connected to your account. +- **A connected CLI instance**: a `kilo remote` CLI running on your own machine. The picker lists the instances currently connected to your account. + +Remote sessions start with the mode and model selected on the new-session screen; older CLI versions that don't accept those fields fall back to their own defaults. By default, the workspace is the CLI's launch directory. Use **Folder** to choose a child folder, including nested folders, before starting. If the CLI cannot list folders, the app explains this and starts in the launch directory instead. In organization context, the new session belongs to that organization. -Remote sessions start with the mode and model selected on the new-session screen; older CLI versions that don't accept those fields fall back to their own defaults. The workspace is always the CLI's own checkout, so there is no repository selection — you type your first prompt in the chat after the session starts. The picker also appears in organization context, where the spawned session is attributed to the organization. +For Cloud Agent sessions, choose a repository from GitHub, GitLab, or, for organizations, Bitbucket. The picker groups repositories by provider and includes **Recently used**. Each provider has its own connection and error messages, so a problem with one does not hide the others. + +## Continuing a finished session + +Open a finished session and tap **Continue** to copy its conversation into a new session. The form starts with the source repository, mode, model, and reasoning variant. Choose **Cloud Agent** or a connected CLI in **Run on**, then tap **Start**. + +The CLI must support importing sessions. If it does not, the app explains why and disables **Start** rather than creating an empty session. An unavailable model or repository also prevents starting until you choose an available option. ## Queueing follow-up messages @@ -62,6 +90,12 @@ The composer stays editable while the agent is working, so you don't have to wai A queued message shows a subtle **Queued** badge on its bubble. The badge clears when the message starts processing or when the queue drains or is cancelled. Queueing works for Cloud Agent sessions and for remote sessions on a connected `kilo remote` CLI instance. +## Voice input + +Use the microphone in the composer to dictate a prompt. By default, the app uses your operating system's speech recognizer. To transcribe through your Kilo account instead, turn on **Gateway transcription** in **Preferences** and choose a transcription model. Your model choice is remembered across launches. + +The selected engine handles the whole dictation; the app does not fall back to the other engine. Tap the microphone to cancel an in-progress transcription. + ## Attachments in remote sessions When you connect the mobile app to a `kilo remote` CLI session, you can share files in both directions. @@ -83,11 +117,13 @@ While the CLI is connected, the agent can deliver a file to your phone with the Open a pull request from a PR link to review it without leaving the app: -- **Overview** — PR state and CI checks at a glance. +- **Overview**: PR state and CI checks at a glance. - **Files** — syntax-highlighted diffs with line-level comments and a file navigator. - **Discussion** — review threads with replies, resolve/unresolve, and reactions. -Comments you leave are collected into a pending review on your device and submitted to GitHub as a single review. When the PR is ready, you can merge it (merge, squash, or rebase), enable or disable auto-merge, or update the branch — all from the app. +Comments you leave on diffs are collected into a pending review on your device and submitted to GitHub as a single review. To post a regular conversation comment instead, tap **Comment on this pull request** at the bottom of **Discussion**. These comments appear directly in the discussion and are not part of a review. + +When the PR is ready, you can merge it (merge, squash, or rebase), enable or disable auto-merge, or update the branch, all from the app. PR review uses your connected GitHub account; the app asks you to connect GitHub if you have not already. @@ -101,6 +137,13 @@ The app shows what each session cost and which models did the work: Cost is recorded when a session closes; sessions that closed before this feature shipped do not show a cost. +## Widgets and live updates + +Add the **Active Agents** widget to track **Needs input**, **Working**, and **Idle** session counts. Tap the widget to open your agents in the app. Background updates keep these surfaces informed while the app is not open; signing out stops updates for that account. + +- **iOS**: widgets are available on the Home Screen and Lock Screen, with a Live Activity on the Lock Screen and Dynamic Island. Compact layouts prioritize agents that need input, then working agents, then idle agents. If Live Activities are disabled, the Agents tab offers a prompt to open device settings. +- **Android**: the resizable Home Screen widget shows **Needs input**, **Working**, and **Idle** counts at every size, including zeros while work is present. An ongoing **Active agents** notification also reports session activity. Supported devices can promote it to a Live Update. If notifications are disabled while work is pending, the app offers a prompt to enable them in device settings. + ## Android App The Android app is available now on Google Play. diff --git a/packages/kilo-docs/pages/code-with-ai/platforms/vscode/index.md b/packages/kilo-docs/pages/code-with-ai/platforms/vscode/index.md index e264507d77a0..a5aa6779311d 100644 --- a/packages/kilo-docs/pages/code-with-ai/platforms/vscode/index.md +++ b/packages/kilo-docs/pages/code-with-ai/platforms/vscode/index.md @@ -31,6 +31,8 @@ Key features include: - **[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 +- **[Notifications](/docs/getting-started/settings/notifications)** - Choose sounds, VS Code alerts, or OS notifications when a session needs attention +- **[Keep Awake](/docs/getting-started/settings/keep-awake)** - Optionally prevent system sleep while agents work in a local VS Code window - **Transcript export:** Save complete local session transcripts as Markdown files - **Sub-Agent Viewer** — Read-only panels for viewing child agent sessions - **Legacy Migration** — Automatic migration wizard for VSCode extension settings @@ -50,13 +52,22 @@ Each agent shows one of these states: An agent can also show **Needs input** when it is waiting for a permission decision or an answer to a question. Open the agent row to inspect its read-only transcript and handle the request in the child-agent view. In the sidebar and Kilo tab, the transcript opens in a read-only editor tab. In Agent Manager, it opens in the right-hand inspector. -Use these controls in the expanded strip: +Use these controls in the background-agent strip: - **Stop** cancels a running agent and its child session. +- **Stop all (N)** cancels every running background agent for the current session without stopping the main session. - **Dismiss** hides one finished agent from the strip. It does not delete the transcript or the agent record. - **Clear finished** hides all agents that are no longer running. - **Continue in background** appears when a foreground subagent is running. It detaches that work so the parent session can continue while the subagent runs in the background. +Stopping the main session with **Stop** or `Esc` ends its current response but leaves background agents running. Use **Stop all (N)** or an individual agent's **Stop** to cancel background work. Completed background results are available when you send the next message or use **Continue**. + +## Follow-up messages + +Send a message while Kilo is working to queue it for processing in order. To revise a queued message, clear the prompt input and click the message's edit icon. This removes it from the queue and restores its text, images, and file references to the input. Sending it again adds it to the end of the queue. + +After stopping a response, leave the input empty and click **Continue** to resume the interrupted turn with its original model and agent, without adding another user message. Continue starts a new model request; it does not undo tool actions already performed and is not available for completed responses. + ## Shared Settings Settings apply across extension surfaces, including the sidebar and Agent Manager. The standalone CLI uses the same `~/.config/kilo/kilo.jsonc` (global) and `./kilo.jsonc` (project) files when used directly. diff --git a/packages/kilo-docs/pages/collaborate/enterprise/sso.md b/packages/kilo-docs/pages/collaborate/enterprise/sso.md index b52f3a8ee643..3397f7481513 100644 --- a/packages/kilo-docs/pages/collaborate/enterprise/sso.md +++ b/packages/kilo-docs/pages/collaborate/enterprise/sso.md @@ -54,6 +54,8 @@ Copy the Service Provider details (Entity ID, ACS URL, and Metadata) from the Wo 1. Set the organization policy and user provisioning settings according to your organization's needs. 2. Configure domain policy and domain verification in WorkOS. +[Verified-domain auto-join](/docs/collaborate/teams/team-management#joining-automatically-with-a-verified-domain) is separate from SSO domain policy. It adds users with a matching email domain to your organization and does not require SSO. + After enabling SSO: - Invite new users with their company email domain. diff --git a/packages/kilo-docs/pages/collaborate/teams/team-management.md b/packages/kilo-docs/pages/collaborate/teams/team-management.md index b78e67d9a60a..7a6049c5031d 100644 --- a/packages/kilo-docs/pages/collaborate/teams/team-management.md +++ b/packages/kilo-docs/pages/collaborate/teams/team-management.md @@ -5,11 +5,11 @@ description: "Add and manage team members in Kilo Code" # Managing Your Team -Every person on the team is an _Owner_ or a _Member_. +Team roles are _Owner_, _Admin_, _Member_, and _Billing Manager_. Owners have full administrative oversight including billing, seat allocation, and model/provider selection. -Only Owners can conduct team management activities. +Owners and Admins can manage the team. Only Owners can grant the Owner role or manage another Owner's membership. Members can use the Kilo Code extension and see data on the team's usage in the [usage dashboard](/docs/collaborate/teams/analytics). @@ -18,11 +18,23 @@ Members can use the Kilo Code extension and see data on the team's usage in the 1. **Navigate to Organization Tab** in your profile page and click on the team you want to manage 2. **Click "Invite Member"** button 3. **Enter the team member's email address** -4. **Select initial role** (Member or Owner) +4. **Select an initial role** from the available options 5. Click **Send Invitation** {% image src="/docs/img/team-management/invite-member.png" alt="invite-member" width="619" caption="invite-member" /%} +## Joining automatically with a verified domain + +An organization Owner or Admin can verify an email domain so users with matching addresses join automatically: + +1. Open the **Verified Domains** card in your Organization dashboard. +2. Add your company's email domain and follow the WorkOS verification steps to prove ownership. +3. Return to **Verified Domains** and click **Check status**. If verification is still pending, check again until the domain shows **Verified**. + +After verification, matching users join as Members when they sign in through the browser, SSO, or the mobile app. They keep their personal account and other organization memberships. Removed members are not automatically re-added, and elevated roles still require an explicit role change. + +Auto-join does not require SSO. See [SSO setup](/docs/collaborate/enterprise/sso) to configure single sign-on separately. + ## Removing Team Members When team members leave: @@ -39,7 +51,7 @@ Promote or demote team members as needed: 1. **Locate team member** in Organization tab 2. **Click role dropdown** next to their name -3. **Select new role** (Member, Owner) +3. **Select a new role** from the available options 4. **Confirm change** 5. **Member receives email notification** diff --git a/packages/kilo-docs/pages/customize/context/context-condensing.md b/packages/kilo-docs/pages/customize/context/context-condensing.md index 9339ed5509f7..a2deb66b02fa 100644 --- a/packages/kilo-docs/pages/customize/context/context-condensing.md +++ b/packages/kilo-docs/pages/customize/context/context-condensing.md @@ -95,7 +95,7 @@ Compaction is configured in your `kilo.jsonc` file: ### Use a different model for compaction -Summarization can use a cheaper or larger-context model than your main agent. Configure a dedicated compaction agent: +Summarization can use a cheaper or larger-context model than your main agent. In VS Code, choose **Compaction model** under **Settings → Models**, or configure a dedicated compaction agent in `kilo.jsonc`: ```jsonc { @@ -109,6 +109,8 @@ Summarization can use a cheaper or larger-context model than your main agent. Co If no compaction agent is set, the current session's model is used. +The selection applies to automatic and manual compaction without changing your chat model. Clear it to use the current session's model again. + ### Environment overrides | Variable | Effect | diff --git a/packages/kilo-docs/pages/customize/custom-instructions.md b/packages/kilo-docs/pages/customize/custom-instructions.md index 064f9be342e3..ce91a7606497 100644 --- a/packages/kilo-docs/pages/customize/custom-instructions.md +++ b/packages/kilo-docs/pages/customize/custom-instructions.md @@ -42,7 +42,7 @@ Place any of these files at your project root to provide project-wide instructio For instructions that apply across all your projects, place an `AGENTS.md` file in your global config directory: - **Kilo:** `~/.config/kilo/AGENTS.md` -- **Claude-compatible:** `~/.claude/CLAUDE.md` +- **Claude-compatible:** `~/.claude/CLAUDE.md`, until [Claude Code Migration](/docs/getting-started/settings#claude-code-migration) has been attempted. Migration ends this global fallback; project-level `CLAUDE.md` files keep working. Project-level instructions are loaded before global instructions and apply to every session. @@ -119,7 +119,7 @@ Place any of these files at your project root to provide project-wide instructio For instructions that apply across all your projects, place an `AGENTS.md` file in your global config directory: - **Kilo:** `~/.config/kilo/AGENTS.md` -- **Claude-compatible:** `~/.claude/CLAUDE.md` +- **Claude-compatible:** `~/.claude/CLAUDE.md`, until [Claude Code Migration](/docs/getting-started/settings#claude-code-migration) has been attempted. Migration ends this global fallback; project-level `CLAUDE.md` files keep working. Project-level instructions are loaded before global instructions and apply to every session. diff --git a/packages/kilo-docs/pages/deploy-secure/security-reviews.md b/packages/kilo-docs/pages/deploy-secure/security-reviews.md index 22c25c8fdc79..9339a83149c3 100644 --- a/packages/kilo-docs/pages/deploy-secure/security-reviews.md +++ b/packages/kilo-docs/pages/deploy-secure/security-reviews.md @@ -375,9 +375,11 @@ SLA notifications require both SLA tracking and SLA notifications to be enabled. ## Notification delivery -Security Agent currently sends notifications only by email. +Security Agent sends notifications by email and mobile push. -Notification kinds: +Mobile push covers analysis completion or failure, and remediation being queued, opening a pull request, failing, being blocked, needing no changes, or being cancelled. These updates go to the owning user for personal accounts or organization owners, subject to their security notification preferences. + +Email notification kinds: | Kind | When eligible | |---|---| @@ -408,6 +410,8 @@ The audit report shows Security Finding activity recorded for an owner during a - `/security-agent/audit-report` - `/organizations/:organizationId/security-agent/audit-report` +You can also open the audit report in the mobile app to see the report period and each finding's recorded activity. + The audit report is based on activity recorded by Kilo. It does not prove that every historical event is present, show repository scan coverage, or provide aggregate SLA compliance. ### Report period and filters @@ -480,7 +484,7 @@ The following capabilities are not yet implemented but are being considered for - GitLab support for Security Agent findings and remediation. - Security Finding sources beyond Dependabot alerts, such as npm audit and SBOM analysis. -- Notification channels beyond email. +- Notification channels beyond email and mobile push. - Historical replay when you enable New-finding Notifications. - Analysis and remediation without queue delays. This work currently runs through queues and can be delayed by account capacity or worker backlog. - Automatic finding updates based on the full remediation PR lifecycle. A finding currently closes only when Dependabot reports it fixed or someone dismisses it. diff --git a/packages/kilo-docs/pages/gateway/models-and-providers.md b/packages/kilo-docs/pages/gateway/models-and-providers.md index e7bcaced7ab1..3fd9c5d570fd 100644 --- a/packages/kilo-docs/pages/gateway/models-and-providers.md +++ b/packages/kilo-docs/pages/gateway/models-and-providers.md @@ -56,16 +56,7 @@ This returns model information including pricing, context window, and supported ### Free models -Several models are available at no cost, subject to rate limits: - -| Model ID | Description | -|---|---| -| `stepfun/step-3.7-flash:free` | StepFun Step 3.7 Flash | -| `poolside/laguna-s-2.1:free` | Poolside Laguna S 2.1 | -| `poolside/laguna-xs-2.1:free` | Poolside Laguna XS 2.1 | -| `nvidia/nemotron-3-ultra-550b-a55b:free` | NVIDIA Nemotron 3 Ultra | -| `tencent/hy3:free` | Tencent Hy3 | -| `openrouter/free` | Best available free model | +Several models are available at no cost, subject to rate limits. Availability changes; check the [live model catalog](https://kilo.ai/models) for current free options and model IDs. Free models are available to both authenticated and anonymous users. Anonymous users are rate-limited to 200 requests per hour per IP address. diff --git a/packages/kilo-docs/pages/getting-started/byok.md b/packages/kilo-docs/pages/getting-started/byok.md index d3915372cba9..f4871f1899bd 100644 --- a/packages/kilo-docs/pages/getting-started/byok.md +++ b/packages/kilo-docs/pages/getting-started/byok.md @@ -24,6 +24,7 @@ Use your provider API key to route matching models through your account: - Anthropic - AWS Bedrock +- Azure Foundry (experimental) - DeepSeek - Fireworks - Google AI Studio @@ -67,7 +68,18 @@ These providers offer coding-focused subscriptions or dedicated endpoints. Bring ### AWS Bedrock configuration -AWS Bedrock requires credentials in a different format than other providers. Instead of a single API key, you must provide your AWS credentials as a JSON object: +AWS Bedrock requires JSON credentials. Use one of these two formats; don't mix fields from both. + +**Bedrock API key:** Generate a key in the AWS Bedrock console and use a region where the key and model are available. Replace the key before it expires. + +```json +{ + "apiKey": "...", + "region": "us-east-1" +} +``` + +**IAM credentials:** ```json { @@ -88,6 +100,32 @@ Your IAM user or role must have the following permissions: - `bedrock:InvokeModel` - `bedrock:InvokeModelWithResponseStream` +### Azure Foundry configuration + +Select **Azure Foundry (experimental)** and enter JSON credentials. Use `resourceName` for the subdomain of your endpoint, such as `my-resource` from `my-resource.openai.azure.com`: + +```json +{ + "apiKey": "...", + "resourceName": "my-resource" +} +``` + +If your deployment names differ from the gateway model IDs, add `modelMappings` to map each model to its Azure deployment: + +```json +{ + "apiKey": "...", + "resourceName": "my-resource", + "modelMappings": [ + { + "gatewayModelSlug": "openai/gpt-5.4-nano", + "customModelId": "my-gpt-5-4-nano-deployment" + } + ] +} +``` + ## How Bring Your Own Key works - When you use the **Kilo Gateway** provider, Kilo checks if there's a BYOK key for the selected model's provider. diff --git a/packages/kilo-docs/pages/getting-started/faq/account-and-integration.md b/packages/kilo-docs/pages/getting-started/faq/account-and-integration.md index d1c2f600fab2..b0d23e203898 100644 --- a/packages/kilo-docs/pages/getting-started/faq/account-and-integration.md +++ b/packages/kilo-docs/pages/getting-started/faq/account-and-integration.md @@ -18,6 +18,16 @@ Your organization will become inaccessible. No charges will be applied. If you have any remaining credits in your organization, you can contact Support to request that they be moved to your personal account. +### How do I delete my account? + +Account deletion is permanent. Cancel active subscriptions before starting. + +1. Open the [profile page](https://app.kilo.ai/profile) and find **Danger Zone**. +2. Click **Delete account**, then **Send confirmation code**. +3. Enter the code sent to your account email and confirm deletion. + +You are signed out when deletion starts and receive an email when it completes. If your organization has disabled your personal account, self-service deletion is unavailable and **Danger Zone** is hidden. Contact your organization admin or Kilo support instead. + ## Integrations ### How do I unlink my GitHub account? diff --git a/packages/kilo-docs/pages/getting-started/settings/auto-approving-actions.md b/packages/kilo-docs/pages/getting-started/settings/auto-approving-actions.md index 070fc59df500..b2780a2474ea 100644 --- a/packages/kilo-docs/pages/getting-started/settings/auto-approving-actions.md +++ b/packages/kilo-docs/pages/getting-started/settings/auto-approving-actions.md @@ -70,7 +70,7 @@ Use the shield button in the prompt controls to toggle runtime auto-approve for Expand **Manage Auto-Approve Rules** to add commands or patterns to your allowed or denied lists. These rules are then appended to the bottom of the approval rules in settings and the config file. -For the `agent_manager` tool, runtime approvals use the requested capability as the pattern: `worktree`, `local`, `overview`, or `prompt`. Prompting an existing managed session always requires an explicit `prompt` approval the first time, even when a broad Agent Manager allow rule already exists. +For the `agent_manager` tool, runtime approvals use the requested capability as the pattern: `worktree`, `local`, `overview`, `prompt`, `stop`, `move`, or `answer`. Prompting, stopping, moving, or answering a managed session requires its own explicit capability approval the first time, even when a broad Agent Manager allow rule already exists. ## MCP Tool Permissions diff --git a/packages/kilo-docs/pages/getting-started/settings/index.md b/packages/kilo-docs/pages/getting-started/settings/index.md index 01ea85c9f070..0edc4f7263b4 100644 --- a/packages/kilo-docs/pages/getting-started/settings/index.md +++ b/packages/kilo-docs/pages/getting-started/settings/index.md @@ -186,6 +186,8 @@ Available experimental settings include: - **LSP integration** - expose language server diagnostics to the agent - **Paste summary** - summarize large clipboard pastes before including them - **Batch tool** - allow the agent to batch multiple tool calls in one step +- **Task Subagent Model Selection** - let you request a different model or reasoning effort for an individual subagent task (off by default) +- **Claude Code Migration** - import supported global Claude Code configuration once (off by default) - **OpenTelemetry** - enable Kilo telemetry and optional OTLP export when configured Advanced options not exposed in the UI can be configured via the `experimental` key in `kilo.jsonc`: @@ -231,3 +233,21 @@ Straightforward tasks can stay solo. Enabling the board does not mean agents are All participants can read the board history, including messages addressed to others. Recipient selection is not a privacy boundary. Peer messages do not grant user approval or change permissions; `HOLD` and `VETO` are advisory, not controls that pause or cancel work. When a main session has board messages, open the **Board** icon in its task header to read them, refresh them, or reset the board. Only the owning top-level session can view or reset its board; child sessions and cloud sessions cannot. Reset clears visible messages only and does not stop agents or clear conversations. See [Kilo Swarm communication](/docs/automate/agent-manager#kilo-swarm-communication) for the board dialog, ownership rules, and recipient-state warnings. + +### Task subagent model selection + +Enable **Task Subagent Model Selection** in **Settings → Experimental**, or set `experimental.task_model_selection` to `true` in `kilo.jsonc`. It is off by default. + +This lets you explicitly request a different model, provider, or reasoning effort for an individual subagent task. The agent keeps normal defaults unless you request an override; it does not select models autonomously for cost or complexity. See [Per-task model selection](/docs/code-with-ai/agents/model-selection#per-task-model-selection-experimental). + +### Claude Code migration + +Enable **Claude Code Migration** in **Settings → Experimental** to import supported global Claude Code configuration on the next backend start. It is off by default and runs once, with no automatic retry. + +The migration imports: + +- Global instructions from `~/.claude/CLAUDE.md` into Kilo's global `AGENTS.md`. +- Standalone skills from `~/.claude/skills/` that contain only a `SKILL.md`. +- Top-level MCP server definitions from `~/.claude.json`, disabled until you enable them. + +Existing Kilo content takes precedence; conflicts and unsupported items are skipped. Your original Claude files are not changed or deleted. After the attempt, Kilo stops loading global Claude instructions and skills as a fallback, but project-level compatibility such as a repository's `CLAUDE.md` is unaffected. A notification reports the outcome and points to a receipt with imported, skipped, and failed items.