- Worktrees: the New Worktree dialog keeps its form when the worktree list changes while open, and a removed worktree leaves the sidebar under every project it was listed in (thanks to @yulia-ivashko).
- **OpenCode Go:** the usage request the extension sends to OpenCode Go now carries the `x-opencode-session` header that OpenCode Go requires from 6 September. Chat traffic already had it, because it goes through OpenCode.
- Message queue: a queued message keeps its attached context, file mentions, and skill; editing it brings them back to the composer.
- Sessions: starting a rename selects the whole title (thanks to @yulia-ivashko).
- Settings: Fixel Text is available as an interface font.
- Thinking effort: picking Default now sticks after a send and across agent or session switches, and a reopened session restores the effort its last message used (thanks to @yulia-ivashko).
- Worktrees: removing a worktree no longer freezes the interface; it runs in the background with a progress toast (thanks to @yulia-ivashko). A worktree created from a branch behind its upstream now fetches first and branches from the remote (thanks to @jtatum).
- Settings: the theme no longer flips when switching sessions across directories, and a theme the extension cannot fully report keeps the current preference (thanks to @kydorn).
- Chat: a turn that OpenCode stopped no longer ends with nothing on screen — what OpenCode reported shows under the last message, and a message an idle session has left unanswered is named as such. The status report (Ctrl/Cmd+Shift+L) now lists the last session errors and rejected sends.
- Composer: pasting a large block of text now offers to attach it as a `pasted-context-N.txt` file instead of flooding the input, with a reference left at the caret; Settings → Chat can make it always attach or always paste inline (thanks to @makeittech).
- Multi-Run groups can now contain more than five models (thanks to @tomzx).
-`/btw` side questions: a btw session now answers the side question instead of carrying on with the parent's plan, and forks at the last completed turn so a reply that is still streaming is never inherited (thanks to @pocharlies).
- Context usage now reports the session cost including everything its subagents spent (thanks to @igorvelho), and undoing or redoing a parent session keeps its subagents at the same point in history (thanks to @alexandrereyes).
- Chat: the text the model writes before asking a question is shown right away instead of staying hidden until the turn ends (thanks to @makeittech).
- Composer: typing three backticks leaves the caret inside the completed code fence, empty inputs keep a visible caret, and platform autocorrect behavior is preserved (thanks to @franzudev, @TTTPOB, and @IbrahimKhan12).
- GitHub Copilot usage now shows a single AI Credits window, matching Copilot's token-based quota (thanks to @jakoss).
- The extension reuses its OpenCode output channel across managed-server restarts instead of creating duplicates (thanks to @TTTPOB).
- Chat scrolling: with "Follow new content while streaming" off, sending while scrolled up leaves the view where it is; a middle-button pan or Shift+Space stops auto-follow like the wheel does (thanks to @pascalandr); PageUp/PageDown in the prompt box no longer shifts the whole panel up.
- Chat no longer crashes or freezes on: very large tool results, which are capped before rendering (thanks to @JSap0914); a code block with JavaScript template strings that sent the highlighter into endless backtracking (thanks to @makeittech); a diff with a truncated header (thanks to @pascalandr); and a draft or recalled message with Windows line endings, which threw "Selection points outside of document" on every visit (thanks to @yulia-ivashko).
- Chat: a session no longer looks frozen after the webview reloads or is opened late — pending permission and question cards come back (thanks to @yangyaofei) — nor after dismissing the agent's questions and sending a new task (thanks to @bashrusakh).
- Chat rendering: question prompts render Markdown (thanks to @pascalandr); bare links next to CJK or full-width punctuation no longer absorb it (thanks to @gaojunran); inline code, chips, and model-picker highlights stay readable in high-contrast themes (thanks to @difagume and @bashrusakh); a completed reasoning block shows in full instead of replaying, the text-selection menu stays inside the viewport, and the sticky user-message header no longer fades over the reply (thanks to @makeittech).
- Chat actions: tool cards with a file path get a quick-open button that opens the file in the editor (thanks to @robertoberto); sending without a selected model explains what is missing (thanks to @rvaldemar); `/init` stays in slash-command autocomplete after the conversation starts (thanks to @Dawnfz-Lenfeng); copying a message keeps Markdown spacing (thanks to @ChangeHow); a manually chosen model survives switching between Build and Plan (thanks to @makeittech).
- Chat: when the turn-ending signal from OpenCode is lost, the working spinner now clears within about a second instead of up to ten (thanks to @makeittech).
- Updating OpenCode no longer fails with a bare "Bad Request": the extension names the release to install and shows OpenCode's own reason when an update is refused (thanks to @mdatsev and @yulia-ivashko).
- "Add Project" now adds the chosen folder to the workspace instead of failing (thanks to @bashrusakh), and the extension starts in the current workspace folder instead of one restored from storage (thanks to @makeittech).
- Sidebar: pending permission and question badges are no longer covered by the hover actions (thanks to @makeittech); worktree branch search hides non-matching branches (thanks to @bashrusakh).
- Settings/Providers: after saving an API key or signing in, the provider no longer shows "Credentials missing" with its models hidden until you switch away and back (thanks to @herjarsa).
- Settings: number fields and selects no longer clip at large font sizes (thanks to @makeittech), and Windows skill paths are classified correctly, so disabled and duplicate skills are hidden as intended (thanks to @Ttungx).
- **Chat context attachments:** diff and file comments, terminal selections, and linked issues/PRs now show in the conversation as compact context cards — source header, captured content behind an expander, your comment below — instead of raw text inside the message.
- **Chat: comment on a reply.** Select text in a chat message and choose Comment to attach that quote with your note to the next message; the selection stays highlighted while you type.
- Composer: hovering a context chip above the input opens a stacked preview of everything attached, where comments can be edited in place or items removed before sending.
- Chat: @ file mentions now rank files and directories together by how well they match, so the file you typed is at the top instead of below unrelated directories. Multi-word queries match in any order, and long paths keep the folder next to the file name visible.
- Search: Ctrl/Cmd+P now matches the whole file path, not just the file name — searching a folder name finds the files inside it.
- Search in dropdowns: searchable pickers (agents, models, providers, branches) now put the best matches first, match multi-word queries in any order, and ignore punctuation (so "gpt4o" finds "gpt-4o").
- Permissions: cards answer to the keyboard with Alt+Enter to allow once, Alt+Shift+Enter to allow always, and Alt+Backspace to deny; the keys are printed on the buttons.
- Keyboard: dropdown menus and pickers answer Ctrl+N/Ctrl+P for down/up, the session switcher opens focused on your current session, and shortcut labels in tooltips and menus show the binding you actually have set (thanks to @ChangeHow).
- Chat: Cmd/Ctrl+Shift+T now cycles through every thinking level offered by the selected model instead of skipping levels after reaching the end (thanks to @nimobeeren).
- Chat: the view no longer stays stuck on its loading screen on slow or remote connections, including code-server behind a reverse proxy (thanks to @VinciYan).
- **/btw side questions:** type `/btw` followed by your question to ask something off-topic in a temporary session forked from the current conversation. The answer streams into a panel above the composer; collapse it, keep it as a full session, or discard it without touching the chat (thanks to @jaygupta17).
- **Skills catalog:** browse curated GitHub skill collections in a card-based catalog with cross-source search and direct links to each skill's repository.
- Settings: the workspace selector on Providers, Agents, MCP, Commands and Skills now only changes what those pages show instead of moving the chat, session list and file tree to another workspace.
- Settings/Integrations: the experimental page now only lists integrations that can be installed; unavailable and Coming soon entries were removed.
- Chat: app links such as `spotify://` now ask for confirmation before opening another app. You can trust an app link type on one device and manage trusted links in Settings.
- While a reply streams, the model status line under the last message now turns into the finished message's info row in place, instead of jumping when the reply completes.
- Settings/General: changing the default model, variant or agent no longer repoints an open chat that already carries a model you picked for it. Chats following the default still switch immediately.
- Settings/Providers: the provider you select no longer jumps to a different one when the chat selection or provider data changes.
- If OpenCode restarts while a response is still running, the chat now stops with an interrupted state and a notification to continue instead of hanging silently (thanks to @sum117).
- Chat: newly sent messages and syntax-highlighted code blocks no longer briefly flicker. Bash output can also grow with its content instead of being cut off.
- **Settings/Integrations:** a new Integrations settings page lists Claude Code, Command Code, and Cursor plugins with install, update, setup, and remove actions, plus Discord and Telegram Coming soon placeholders.
- Usage: Command Code plan limits now appear in the Usage page and work status panel.
- Usage: the context usage readout in the chat header now also shows the session's cost in its tooltip (thanks to @YunFeng0817).
- Attachments: extracted Office and OpenDocument content is now capped and presented more compactly, preventing large documents and their images from overwhelming the message context.
- Skills Catalog: the source is now named ClawHub instead of "ClawdHub" (thanks to @makeittech).
- **Chat:** an open conversation no longer keeps re-coloring the same code blocks in the background, preventing high CPU use while the chat is idle (thanks to @makeittech).
- Settings: OpenChamber no longer replaces a full OpenCode config with an empty `$schema`-only stub when the file uses JSON5-style unquoted keys; Settings changes now fail instead of wiping plugins, MCP servers, and providers (thanks to @makeittech).
- Chat: new chats no longer start against a deleted last worktree directory; they fall back to the active project instead of saving the first message and never starting.
- Chat: typing with Chinese, Japanese, or Korean input methods no longer interrupts composition or jumps the cursor to the end of the composer (thanks to @makeittech).
- The context usage readout no longer climbs over 100% after turns with many tool calls and no longer jumps when reopening an older session; it now shows what the window actually holds (thanks to @pocharlies).
- Projects: project names now match the folder name exactly, so `.ssh` and `opencode-claude` are no longer shown as `.Ssh` and `Opencode Claude`; names you renamed yourself are kept.
- The model selection menu no longer shows white text on a white highlight when a high-contrast theme is active, so the hovered or selected model stays legible (thanks to @bashrusakh).
- **Chat:** new messages now remain at the end of the conversation instead of jumping before older messages after the message ID sequence rolls over; history loading, revert, and redo follow the same chronological order.
- Sessions: switching projects now selects a session owned by the new workspace, and a message already being prepared stays with the session where it was submitted instead of being rerouted by a later project switch (thanks to @makeittech).
- Chat: the Focus Chat command and Add to Context action now place the cursor in the chat input so typing can continue immediately.
### Improvements
- Chat images: completed assistant replies now collect Markdown images into a compact gallery with thumbnails and full-screen previews, including workspace-local images across multi-root workspaces (thanks to @ChangeHow).
- Chat: typing `!` to enter shell mode no longer inserts the trigger into the command or moves the caret to the wrong side of it (thanks to @RyderAsKing).
- Chat: line numbers with three or more digits no longer wrap in code blocks (thanks to @ChangeHow).
- Chat: messages submitted before switching sessions stay with the session and workspace they were sent from, and are cancelled rather than crossing into a different runtime (thanks to @Wsyjq).
- Chat: shell command output is expanded by default, and adding a message to context returns focus to the composer (thanks to @pascalandr, @makeittech).
- **Settings:** OpenCode configuration changes now accumulate behind a single Apply & Restart action instead of restarting OpenCode after every edit; the confirmation warns when active chats will be stopped (thanks to @makeittech).
- Chat: queued messages no longer send into a response that is still streaming, and tool cards left running by an interrupted response settle instead of remaining stuck (thanks to @makeittech).
- Chat: fresh messages no longer replay their entry animation after they have already been shown (thanks to @makeittech).
- MCP: authorization now handles browser callbacks more reliably, settings distinguish available and unavailable servers more clearly, and failed connections expose a retry action.
- Settings: rapid edits to notification templates no longer overwrite one another, and the collapsed-user-message preference now persists correctly (thanks to @AmanTahiliani, @pascalandr).
- **Providers:** signing in to an OAuth-only provider (such as Cursor) now completes in the browser — the login is stored and the provider updates instead of remaining signed out. OAuth-only providers show a Connect flow instead of an API key form, and their models stay hidden until you are signed in.
- Providers: sign-ins that need extra details (such as GitHub Copilot Enterprise) now ask for them before opening the browser, and device codes come with a working copy button.
- Chat: a manually chosen model now stays selected after a delegated subtask finishes, instead of reverting to the agent's default model.
- **Sessions:** archived sessions can now be restored to the active list — from the sidebar context menu, the archived-sessions page, or the bulk-selection bar (thanks to @makeittech).
- **Providers:** custom OpenAI-compatible providers can now be added and edited from Settings, including their endpoint, models, credentials, headers, and configuration scope (thanks to @makeittech).
- UI/Localization: added German interface translations (thanks to @SGD-DEV).
- Chat/Tools: Bash output now applies terminal control characters and strips ANSI formatting, preventing progress output and rewritten lines from appearing as raw escape sequences (thanks to @catan271).
- Settings/Skills: repository-local `.agents/skills` now appear for the active workspace (thanks to @makeittech).
- Settings/Skills: renaming a skill now preserves its instructions and supporting files; only skills in locations OpenChamber can safely rename show the action (thanks to @makeittech).
- Usage: Kimi for Coding now calculates usage correctly when the provider reports either used or remaining quota (thanks to @makeittech).
- Chat: clicking an apply_patch tool result now opens each changed file at its correct path instead of always opening the first file (thanks to @nabsiddiqui).
- **OpenCode updates:** the extension can now update a managed OpenCode installation and restart it automatically; externally managed servers are left unchanged (thanks to @yulia-ivashko).
- Chat: slash-command starters now include text already entered in the draft as command arguments.
- Usage: OpenAI business-account Codex usage now shows the configured spend limit (thanks to @jrandiny).
### Fixes
- **Chat tools:** Bash tool cards now show output before a command finishes, keep it in a fixed-height pane, and follow new lines until you scroll away. Long-running commands no longer remain at a 300-second duration, and their timers continue until they finish.
- Chat: selecting text from Markdown code blocks now preserves the code fences, language, and surrounding block structure when adding it to the composer or starting a new session (thanks to @ChangeHow).
- Settings: added an option to hide starter suggestions on the new-session screen.
- **Chat composer:** prompts now render Markdown emphasis, attention lines, file and agent mentions, slash commands, snippets, attachment citations, and `~path` references directly while you type. File mentions can be edited in place.
- Sessions: the sidebar now uses clearer project zones and single-line session rows, with folders shown after ungrouped sessions and archived sessions kept in their existing workspace sections.
- Chat/Permissions: sending a message while a permission prompt is open now denies pending requests in the session and its subagents, then queues the message for the next turn (thanks to @tomzx).
- **Context panel:** a new surface rail brings Changes, pull requests, files, terminal, notes, plans, previews, and side chats into one resizable panel. The pull-request surface now shows live checks and comments, and can attach failed checks or comments to a chat draft.
- Sessions: fixed missing activity-based sorting; chats now move in the list when they start or finish instead of repeatedly reordering while a response streams.
- Chat: code blocks no longer shift line layout or merge adjacent text while rendering, and copied code keeps its original text (thanks to @ChangeHow).
- Usage: added Crof and NeuralWatt quota tracking, including the missing Crof switch arm that previously fell back to "Unsupported provider" (thanks to @kydorn).
- **Chat attachments:** added Office and OpenDocument files (`.docx`, `.pptx`, `.xlsx`, `.odt`, `.odp`, and `.ods`), with readable text and supported embedded images extracted before sending. Attachments also support more source-code formats, notebooks, HAR files with credentials and cookies removed, SVG and Draw.io files, and HEIC/HEIF images; the composer warns when the selected model may ignore an attachment type.
- Chat/Tools: attachments returned by plugin and custom tools remain visible after streaming and refreshes, with the same image previews and file chips as chat attachments (thanks to @FrostiDrinks).
- Sessions: workspace groups now default to manual ordering instead of recent-activity order; explicit sorting choices remain unchanged.
- **Performance:** opening and switching sessions now prioritizes the selected and visible chats in large workspaces. Failed refreshes keep the existing session list, parent sessions no longer disappear when their sub-sessions load first, and session data no longer crosses between workspaces or extension views.
- Chat: assistant turns now show model, agent, thinking level, duration, and time together in the footer, and replies separated by hidden system or subagent prompts display as one continuous turn. The working indicator shows the model actually producing the active response, streaming at the bottom no longer jitters, and new user messages finish their entry animation instead of snapping into place.
- Cursor: opening a chat no longer crashes when the editor webview does not expose its usual messaging APIs, and disposed editor tabs no longer receive late streaming messages (thanks to @makeittech).
- Startup: the active workspace is now detected before startup state is restored, preventing projects outside the editor workspace from replacing it.
- Agent Manager: worktree creation now recovers when an earlier Git operation left the repository locked, and removing a worktree while setup is still running no longer brings it back.
- Sessions: new drafts and sessions now stay with the workspace selected in the sidebar, including multi-root and nested workspace setups (thanks to @bashrusakh).
- **Settings:** pages now use a consistent responsive layout, navigation is grouped by area, and save failures are shown in the page header. Agent tool permissions now distinguish inherited and explicit rules and show session-granted rules separately (thanks to @makeittech).
- Permissions: per-session auto-accept works again, persists across extension restarts, and applies to subagent sessions while an OpenChamber view is open.
- Chat: shell-mode command cards now update their status and output while the command runs, with syntax highlighting for the command and output.
- Chat: task (subagents) now track the correct subagent when several run at once, preventing one subagent's activity or "Open subtask" action from pointing to another session.
- **Performance:** large workspace session lists no longer regroup every session while chats stream. Opening a long chat after an empty or aborted agent turn also no longer repeatedly loads larger portions of its history.
- Chat: sessions with an active [goal](https://docs.openchamber.dev/session-goals/) (started from the web or desktop app) show the goal strip with its live status above the composer.
- Chat: code blocks highlight correctly again — the webview's security policy was blocking the syntax highlighter (thanks to @bashrusakh).
- Chat: queued messages now send when the session is already idle instead of waiting forever (thanks to @bashrusakh).
- Chat: code blocks can show line numbers that stay aligned while streaming, and a new Wrap Code Block Lines setting controls long-line wrapping.
- Editor Integration: the "Add to Context" command and active-editor pin-selection now use workspace-relative filenames so the model reads the correct file when names collide (thanks to @Catan).
- Chat/Tools: every tool call now expands to show its input, result, and errors, including MCP, plugin, and custom tools; Read and Skill stay compact links to their files. JSON results now offer navigable summary, tree, and raw views.
- Chat/Tools: expanded file-edit and patch results include per-file buttons to open the diff or jump to the first changed line in the editor.
- Chat/Thinking: reasoning parts stay separate and in chronological order instead of merging into one block, and collapsed previews no longer show empty trailing HTML comments.
- Chat: if sending a message times out or loses the connection after OpenCode accepted it, the extension now keeps the sent message instead of rolling it back as failed.
- Startup: on Windows, OpenCode installed via npm now launches from paths with spaces (such as C:\Program Files\nodejs), a binary path pasted with surrounding quotes into the Opencode Binary setting works, and discovery also checks the system-wide npm prefix and Scoop's shims.
- Chat: the stop button now aborts sessions running in a different project or worktree than the currently open one — previously those aborts silently did nothing.
- Startup: the extension no longer picks OpenCode desktop app installs when looking for the standalone OpenCode CLI.
- Chat: fixed edge cases where late-loading tool content, subagent content, or streaming Thinking blocks could pull the conversation away from the latest message or fight manual scrolling.
- Chat: embedded JSON examples in messages no longer render as generated-result cards.
- Sync: chat state now recovers after idle reconnects instead of leaving sessions stuck in a stale busy state.
- Chat: a new Follow-up behavior setting controls what happens when you press Enter on a message while the agent is still responding — Steer inserts it into the agent's current turn, or Queue holds it until the turn finishes. Replaces the previous queue-mode toggle (thanks to @bashrusakh).
- Sync: a connected but quiet session (for example an agent running a long tool call) no longer triggers repeated background refreshes every ~15 seconds (thanks to @tomzx).
- Providers: the Add provider form stays open while provider data refreshes or a model is picked in the background, instead of snapping back to an existing provider.
- Chat: with tool calls (such as Bash and Edit) shown expanded by default, scrolling no longer twitches, and slow scrolling no longer jumps past several messages.
- Models: the model picker now remembers provider group expansion and custom ordering, and Shift+Delete removes a recent model from recents (thanks to @makeittech).
- Chat: sending a message now closes an open question prompt instead of leaving stale question UI in the composer (thanks to @tomzx).
- Chat: conversations pinned to the bottom no longer jiggle or double-scroll after sending, and revisiting older sessions snaps to the latest message without a smooth-scroll delay.
- Agents: agent edits against an external OpenCode server no longer show a saved-state update when the save did not succeed (thanks to @makeittech).
- Providers: the add-provider form no longer loses the selected provider during background provider refreshes (thanks to @IbrahimKhan12).
- Chat: selecting a user-installed skill from the slash command menu now invokes the skill instead of inserting the skill name as plain text (thanks to @IbrahimKhan12).
- Chat: code blocks in user messages now preserve characters like `<` and `->` instead of escaping them inside the code block (thanks to @bashrusakh).
- Chat: Arrow Up opens prompt history again when the cursor is at the start of the composer.
- Agents: agent settings now include thinking variant, temperature, and top-p controls, and clearing temperature or top-p now removes the override (thanks to @bashrusakh).
- Settings/Models: per-model visibility and sibling model selections now stay saved after changes (thanks to @attilaszasz).
- Settings: font size and padding preferences now apply inside the extension webview (thanks to @Sin991114).
- Settings/Skills: the skills catalog refreshes after catalog settings change (thanks to @gokulkgm).
- Usage: MiniMax M3 and Token Plan usage now handle the provider's latest API response format (thanks to @baruchvitorino).
- Chat: pasted text containing `@` no longer opens file mention autocomplete unexpectedly (thanks to @charpeni).
- Chat: switching sessions and loading older messages no longer causes the conversation to jump backward or oscillate around the current scroll position (thanks to @herjarsa).
- Sessions: new sessions now stay attached to the selected workspace directory instead of sometimes appearing under a stale project (thanks to @bashrusakh).
- Sessions: pinned sessions and folder rows no longer disappear from the sidebar after an empty session-list refresh (thanks to @bashrusakh).
- Chat: the end of a streamed response is no longer occasionally cut off — messages now always settle on their complete text (thanks to @IbrahimKhan12).
- Startup: the extension starts faster by no longer waiting on default OpenCode config, while your manual and per-directory model selections are preserved.
- Settings/MCP: importing MCP snippets from OpenCode config works again (thanks to @youzini).
### Fixes
- Startup: the extension opens faster — recent sessions, models, providers, and projects appear instantly from cache and refresh in the background, and the loading screen no longer lingers after the interface is ready.
- Startup: requests made while OpenCode is still starting now wait briefly for it to become ready instead of failing, and if OpenCode fails to start the error now includes what it reported.
- Chat/Input: tab-completing a mention no longer changes the selected agent (thanks to @Quat3rnion).
- Chat: LSP tool calls now show the operation, file, and cursor position more clearly, and JSON tool output can be toggled between formatted and raw views or copied.
- Chat: streaming messages now appear correctly after startup, and activity/status rows show for the active session.
- Sessions: session titles update from live session events, and the extension now consistently loads all existing OpenCode sessions.
- Settings/OpenCode: OpenCode CLI path, update-notification preference, keyboard shortcuts, and protected-session settings now stay saved after changes.
- UI/Time: the time-format preference now applies to chat timestamps, usage reset times, scheduled tasks, passkeys, and usage last-updated times.
- Chat: added customizable draft welcome starters from commands and skills, including guided commands for catch-up, debugging, exploration, and approach comparison.
- Chat: assistant answers now have a dialog for starting a new session from that answer.
- Chat/Input: queued messages no longer auto-send before the active session is ready, and thinking-variant choices are preserved for generated messages.
- Chat/Input: selecting an agent now switches to that agent's configured model, and malformed tool diffs no longer break chat rendering (thanks to @Adrian-Eckardt).
- Chat/Input: pending image attachments now show previews, sent image attachments can be cited from assistant messages, and markdown source mode highlights formatting while you type.
- Chat: queued messages now send to the session they were queued from, even if you switch sessions before they are sent.
- Settings/Skills: installed skills are discovered more accurately, skill files opened from tool messages now load correctly, and snippet names keep their canonical casing (thanks to @jkker, @isanchez404).
- Chat/UI: chats keep following the latest response after final task summaries, activity reasoning no longer flashes before settling, and assistant timestamps stay visible on narrow layouts.
- Chat: session editor tabs now update their title to match the session title.
- Sessions: session rows now include an action to open a chat as editor tab.
- Notifications: completion, question, and permission notifications now use the extension notification settings and shows as multi-platforn native notifications.
- Chat: question cards now include copy buttons for Markdown and JSON (thanks to @robertoberto).
- Chat/Input: queued messages now auto-send one at a time in FIFO order, and model/agent selections persist across reloads (thanks to @lyxxx708, @chutastic).
- Chat/Performance: virtualized more timeline content, deferred heavy tool output, and improved scroll-to-bottom behavior.
- Editor Integration: chat commands now wait for the sidebar webview before sending selections, file mentions, explain prompts, or improve prompts.
- Reliability/Streaming: extension SSE reconnect delays now abort cleanly and disposed chat webviews clean up their live streams.
- Agent Manager: settings changes now sync into Agent Manager views immediately.
- Sessions: archived-session bulk delete now works reliably from the extension sidebar (thanks to @jjdubski).
- Git: deduplicated lightweight and full status refreshes separately, preventing stale or mismatched Git updates during background polling (thanks to @isanchez404).
- Reliability/Sync: reduced stale and duplicate live-state updates across request arrays, retry metadata, streaming indicators, and session status events, cutting unnecessary rerenders and stuck activity states during long-running chats (thanks to @isanchez404).
- Chat/Reliability: added smaller polish fixes for text-selection cleanup (thanks to @isanchez404).
- Terminal: improved Android tablet keyboard handling, including control-key shortcuts, and kept app shortcuts from stealing focus while typing in the terminal (thanks to @Dav1dch).
- Startup/Reliability: configured OpenCode CLI paths are now validated before managed startup, with clearer errors for missing, non-executable, or app-bundle paths.
- Chat/Scrolling: preserved per-session scroll position and kept generated prompts scrolled into view (thanks to @jwcrystal).
- Settings/UI: improved settings scrolling and empty states (thanks to @Yabuku-xD).
- GitHub/MCP: improved fork-aware issue and pull-request listing, PR status handling, and remote MCP header handling (thanks to @corrm, @ricautomation).
- Chat/Streaming: reduced text flicker, preserved first chunks reliably, and kept long agent sessions from hanging during active responses (thanks to @pasta-paul).
- Models/Providers: fixed slash-containing model IDs, refreshed model metadata after OpenCode restarts, and added safer concurrency controls for sessions sharing the same provider (thanks to @yart, @Yabuku-xD).
- Chat/Models: unified the model picker with a cleaner selection flow (thanks to @daveotero).
- Chat/UI: improved split-response action placement, error-message alignment, tab close affordances, and overscroll behavior.
- Windows: normalized drive-letter paths in extension webviews (thanks to @sdunfeng).
- Reliability/Startup: hardened managed OpenCode startup, preserved shell PATH reliably, and improved stream/proxy recovery with heartbeat support (thanks to @An-jinu).
- Sessions/Sidebar: fixed stale session, folder, project, and worktree state after mutations, and polished pinned-session indicators (thanks to @corrm, @Yabuku-xD).
- UI/Localization: added translated interface strings for Spanish, Brazilian Portuguese, Ukrainian, and Simplified Chinese, with language selection available in extension settings.
- Chat/Tasks: todo list progress now updates live as task status changes, and task/model status hints are steady during active runs (thanks to @Yabuku-xD).
- Chat/Performance: improved cold start and streaming smoothness with lazy-loaded heavy components, chunk-load recovery, and lower re-render churn in long sessions (thanks to @Yabuku-xD).
- Chat/Commands: added `/summary` slash command for a non-destructive session summary - optional topic hint after the command focuses the output, and the prompt is customizable under Settings: Magic Prompts.
- Chat/Files: you can now drag files and folders from the file tree into chat, with improved `@folder` autocomplete when building prompt context (thanks to @youfch).
- Sync/Performance: optimized multi-session streaming with per-directory queues, event coalescing, and parts-gap recovery for steady live updates in busy workspaces (thanks to @jwcrystal).
- Sessions/UI: kept active sessions visible in Recent, auto-expanded parent groups for subagent sessions, and hid empty archived/folder sections (thanks to @jwcrystal).
- Reliability/Windows: normalized workspace path handling in SSE event lookup to keep live session updates working consistently on Windows (thanks to @widipa).
- Chat/Performance: reduced streaming re-render fanout and status-row churn for smooth long responses in the editor panel.
- Chat/Tool Output: LSP diagnostics now render directly in tool output (thanks to @yulia-ivashko).
- Files/Markdown: added filesystem stat support in the extension bridge to validate markdown targets reliably before file handling flows (thanks to @geekifan).
- Chat/Models: added arrow-key navigation for thinking-mode selection in model controls (thanks to @daveotero).
- Chat/JSON: added an interactive JSON tree viewer with collapse/expand controls and richer color cues for easier inspection of large structured outputs (thanks to @nguyenngothuong).
- Reliability: managed server startup now imports login-shell environment values and normalizes Windows paths to reduce session-loading mismatches and proxy-related connection issues.
- Navigation/UI: refreshed the extension shell with a redesigned sidebar, clearer hierarchy, and cleaner session grouping.
- Chat/Performance: reduced streaming overhead and update churn for smooth long responses, steady activity rendering, and fewer UI stalls in heavy sessions.
- Chat: improved follow-to-latest behavior and timeline stability.
- Reliability/Windows: normalized workspace drive-letter handling and hid background process windows to reduce startup/session mismatches (thanks to @zerone0x).
- Chat: completed a turn-based render pipeline with steady streaming, smooth auto-follow, and more stable activity/tool progress behavior during long responses.
- Chat: improved code block readability with cleaner header actions, restored horizontal scrolling, and themed highlighting in markdown and tool output (thanks to @nelsonPires5).
- Usage: added MiniMax coding-plan quota provider support (thanks to @nzlov).
- Usage: added Ollama Cloud quota provider support (thanks to @iamhenry).
### Improvements
- Chat: improved long-session performance with virtualized message rendering, smooth scrolling, and more stable behavior in large histories (thanks to @shekohex).
- Chat: added fullscreen Mermaid preview, improved default thinking-variant persistence, and hardened file-preview safety checks for a more predictable message experience (thanks to @yulia-ivashko).
- Settings: added an MCP config manager UI to simplify editing and validating MCP server configuration (thanks to @nguyenngothuong).
- Chat: added C, C++, and Go language support for syntax-aware rendering in code-heavy workflows (thanks to @fomenks).
- Settings: grouped agents/skills navigation by subfolder to make larger setups easier to manage (thanks to @nguyenngothuong).
- Chat: draft text now persists per session, and the input supports an expanded focus mode for longer prompts (thanks to @nguyenngothuong).
- Sessions: expanded folder management with subfolders, cleaner organization controls, and clearer delete confirmations (thanks to @nguyenngothuong).
- Chat Activity: improved Structured Output tool rendering with dedicated title/icon, clearer result descriptions, and more reliable detailed expansion defaults.
- Sessions: added custom folders to group chat sessions, with move/rename/delete flows and persisted collapse state per project (thanks to @nguyenngothuong).
- Notifications: improved agent progress notifications and permission handling to reduce noisy prompts during active runs (thanks to @nguyenngothuong).
- Reliability: refactored OpenCode config/auth integration into domain modules for steady provider auth and command loading flows (thanks to @nelsonPires5).
- Skills: upgraded editing/install UX with better code editing, syntax-aware related files, and clearer location targeting across user/project .opencode and .agents scopes.
- Reliability: managed runtime startup now rotates secure auth credentials and hardens API proxy auth forwarding for safer local connections (thanks to @yulia-ivashko).
- Reliability: extension startup/shutdown handling is more predictable.
- Usage: added per-model quota groups in the header and fixed provider dropdown scrolling for easier usage tracking (thanks to @nelsonPires5, @gsxdsm).
- Reliability: fixed OpenCode auth pass-through/proxy behavior to reduce failed extension requests (thanks to @gsxdsm).
- Added multi-provider quota dashboard in settings to monitor API usage across OpenAI, Google, and z.ai with auto-refresh support (thanks to @nelsonPires5).
- Added Agent Manager panel to run the same prompt across up to 5 models in parallel (thanks to @wienans).
- Added permission prompt UI for tools configured with "ask" in opencode.json, showing requested patterns and "Always Allow" options (thanks to @aptdnfapt).
- Added "Open subAgent session" button on task tool outputs to quickly navigate to child sessions (thanks to @aptdnfapt).
- Added timeline dialog (`/timeline` command or Cmd/Ctrl+T) for navigating, reverting, and forking from any point in the conversation (thanks to @aptdnfapt).
- Added `/undo` and `/redo` commands for reverting and restoring messages in a session (thanks to @aptdnfapt).
- Added fork button on user messages to create a new session from any point (thanks to @aptdnfapt).
- Default location is now the right secondary sidebar in VS Code, and the left activity bar in Cursor/Windsurf; navigation moved into the title bar (thanks to @wienans).
- Sidebar: improved readability for sticky headers with a dynamic background.
