Cline

Added

• Restore the "View Changes" button on completion rows, backed by SDK checkpoints, so you can review everything a task touched from the completion card.
• Bring back a copy button on turn-final response rows.
• Support pre-registered OAuth clients for remote MCP servers, for setups where dynamic client registration isn't available.

Changed

• Fade the "View Changes" button until changes since the last message are confirmed, and hide it entirely when there is nothing to show.
• Centralize plugin settings and contributions, with host-aware snapshots and atomic plugin toggles.
• Carry execution context in scheduled run reports — readable headers, schedule metadata, durations, and lifecycle error details.

Fixed

• Preserve prompts queued during a turn when that turn is interrupted: they survive aborts, are drained after a turn aborts itself, and the stop is surfaced instead of the queue being silently dropped.
• Keep session context durable across aborts and hub restarts, so an interrupted session resumes with the state it had.
• Settle the turn phase when a mode switch aborts a running turn.

• Added cline mcp uninstall for removing an installed MCP server
• Schedules now reuse your saved provider settings instead of needing provider configuration of their own
• Queued messages are legible on light-theme terminals — they were previously rendered in a color that washed out against a light background
• MCP tool results render as readable text in the TUI instead of escaped JSON, and binary payloads survive being expanded instead of being mangled
• Malformed tool input/output payloads no longer break rendering — the formatters degrade gracefully instead of throwing
• Prompts queued during a turn now survive being interrupted: they are preserved across aborts, drained after a turn aborts itself, and the stop is surfaced instead of leaving the queue silently dropped (from SDK v0.0.72)
• Session context stays durable across aborts and hub restarts, so an interrupted session resumes with the state it had (from SDK v0.0.72)
• A hung MCP server no longer takes down session creation, and stdio servers that were never configured get a 30-second initialize budget instead of blocking indefinitely (from SDK v0.0.72)

• Prompts queued during a turn now survive being interrupted: they are preserved across user-initiated aborts, drained after a turn aborts itself, and edits made to the queue inside the abort window are applied rather than lost. Stopping a session now has consistent full-stop semantics across hosts
• Session context stays durable across aborts and hub restarts, so an interrupted session resumes with the state it had rather than a reset one
• Queued turns that fail are now reported as run.failed instead of completing silently
• A hung MCP server no longer takes down session creation, and stdio servers that were never configured get a 30-second initialize budget instead of blocking indefinitely
• Remote SSE MCP servers now surface an OAuth authorization prompt on a 401 instead of failing outright, and remote MCP supports pre-registered OAuth clients for setups where dynamic client registration isn't available
• LiteLLM requests route through Chat Completions instead of the Responses API, fixing calls against LiteLLM proxies
• Network interruptions that happen mid-stream but before any model output are retried instead of failing the turn

• Remote MCP servers can now authenticate with OAuth from Settings → MCP — authorize a server, see its auth status, and cancel or retry a pending authorization. Servers that require a pre-registered OAuth client (client ID/secret) instead of dynamic registration are now supported, and stored tokens are invalidated when a server's client configuration changes.
• MCP errors are now shown on the individual server rather than as a page-level error, and a server with invalid configuration is surfaced with its error instead of silently disappearing from the list.
• Failed turns no longer fail silently. Sending a message with no model credentials — or any queued turn that fails — now shows an error in the transcript, enriched with the underlying cause and a pointer to Settings → Models.
• Fixed the first message of a chat (and some queued messages) rendering twice.
• Fixed the composer getting stuck on "Agent is working…" after a turn already finished.
• New "Connect a model" notice on the welcome screen when no provider has credentials, with one click to onboarding or model settings. It reacts live as you add credentials, and correctly recognizes Bedrock/Vertex and keyless local

Added

• Recognize Chutes as a provider.
• Show skills alongside workflows in the slash command menu, and disambiguate commands that share a name instead of letting one shadow the other.

Changed

• Remove model-initiated plan-to-act switching. Switching out of plan mode is now driven by you, not by the model deciding mid-turn.
• Hard-block file-editing shell commands in plan mode instead of relying on prompting alone. Read-only investigation still works, but file manipulation, in-place editors, redirection to files, mutating git subcommands, and package installs are refused.

Fixed

• Stop treating a turn that completes with a plan as a failed turn when a plan-blocked command was its only tool call. The turn no longer ends in the error state with a Retry footer, and toggling to Act correctly re-runs the presented plan instead of appearing to do nothing.
• Show tool paths relative to the workspace in the chat view instead of absolute paths.
• Reset pending attachments when starting a new task, so images from the previous task no longer carry over.
• Surface a clear error when the selected provider has no API key configured, instead of a generic failure.

• Cline Code now ships as a single universal macOS download that runs natively on both Apple Silicon and Intel — no more picking the right architecture. Existing per-architecture installs migrate to it automatically on their next update.
• Session history can now be filtered by where a session came from — Desktop, CLI, extension, or scheduled — from a new filter control in the sidebar.
• The composer now shows a token usage ring for the active model's context window, with cumulative cost, and it changes color as you approach the limit.
• Skills now appear in the slash command menu alongside workflows, and commands that share a name are disambiguated instead of shadowing each other.
• Installed plugins now show their real package names instead of all appearing as "index".
• The agent header can be dragged to move the window again, including on read-only titles.
• Chat message actions (copy, fork, edit, restore) no longer collide with the descenders of the message's last line.
• Application errors are now reported in diagnostics, and the packaged app's telemetry configuration is baked into the sidecar at build time — previously the packaged build shipped with it empty, so no

• Added user-selectable color themes to the interactive TUI. Pick one with /theme, the command palette, or the Theme row in /settings — the picker previews each theme live. Built-in themes are Auto (terminal-adaptive, the default), Cline Dark, Cline Light, Tokyo Night, Gruvbox Dark, Nord, Dracula, Catppuccin Mocha, One Dark, Solarized Dark, and Solarized Light. Named themes paint the background, foreground, accents, syntax highlighting, and diff colors, and CLINE_THEME overrides the persisted choice at startup
• The git branch shown below the prompt now updates when you switch branches from another terminal or your editor, instead of showing whatever was checked out when the TUI started
• Telegram slash commands such as /clear now reach the connector command host — the Telegram library was intercepting them and they were silently dropped
• Racing connector launches no longer collide: an instance is claimed before it opens socket mode, the hub supervises connector processes, and doctor/connect skip connectors that are already starting. Connector tools are also enabled by default, and the Slack greeting is no longer replayed on reconnect

• Plan mode now hard-blocks file-editing shell commands instead of relying on prompting alone — run_commands stays available for read-only investigation, but file-manipulation commands, in-place editors (sed -i, perl -i), redirection to files, mutating git subcommands, package installs, and nested command strings (sh -c, eval, sudo, xargs) are rejected with a tool error, on Windows and PowerShell too
• Context-window overflow errors are now detected and recovered from instead of surfacing as raw unclassified provider errors: the runtime force-compacts with a deterministic strategy that needs no extra LLM call and retries the run once, and terminal cases (nothing left to compact, a retry that still overflows) fail with an actionable message
• Sessions now record how they began — a new mode on StartSessionInput (user, automation, subagent, team) alongside source — and root-session persistence is lazy: starting a runtime allocates the session id in memory without writing a database row, so closing it before any user turn no longer leaves an empty history entry

Fixed

• Stop the two bundles of the combined rollout package from invalidating each other's Cline account session. A still-open legacy window that refreshed its token after the machine was promoted to the new extension would consume the shared refresh token, producing spurious "Unauthorized" / re-authenticate prompts and unexpected sign-outs. Promoted legacy windows now keep working on their current session and offer a one-time Reload Window prompt instead.
• Fall back to the default Cline model when migrating a setup that references a model id the new extension doesn't recognize, instead of leaving the provider unconfigured.
• Restore reliable checkpoints: checkpoints are created consistently, and restoring one now rewinds the whole workspace rather than a subset of files.
• Keep settings edits that are made before the provider config finishes loading — base URLs, API keys, and the Qwen/Moonshot API line are no longer silently discarded.
• Stop losing keystrokes in custom base URL fields, and keep the custom URL checkbox state after a failed clear.
• Use the AskSage custom API URL at inference time instead of ignoring it.

• Edit any earlier message in a conversation — the app forks the session at that point, rewinds the workspace to that run's checkpoint, and re-runs from your edited prompt. Restores are transactional and workspace-atomic, so a failed restore won't leave you half-rewound.
• Fixed long-running chat turns timing out mid-response.
• Checkpoints are now created reliably — including after a restart, after compaction, and on the first turn of a resumed session — and restoring one rewinds the whole workspace, not just the conversation.
• Fixed checkpoint restores failing after you closed and reopened a session.
• Reasoning controls (effort, thinking budget, on/off) now come from the shared model catalog, so each model gets exactly the reasoning options it actually supports instead of provider-specific guesses.
• Errors from upstream providers forwarded through the gateway now show the real message (e.g. "This model's maximum context length is 40960 tokens…") instead of a raw validation dump or [object Object].
• Ollama: empty responses are retried automatically, and the response-start timeout is raised to 5 minutes so cold model loads don't error out.

• /undo works again once the agent has used tools — the checkpoint picker counted tool results as user turns, so restore aborted with "Could not find user message for run N"
• Checkpoints are actually created again; a run-boundary regression meant none were ever recorded in the CLI (from SDK v0.0.69)
• Checkpoint restore is now a full workspace rewind: files Cline created during the task come back at their checkpoint-time content and files created after the checkpoint are removed, while .gitignored paths (build output, node_modules, .env) are left alone (from SDK v0.0.69)
• After a restore, the rewound message is prefilled as plain text instead of the raw <user_input mode="act"> envelope
• Ollama's response-start timeout is now 5 minutes instead of 30 seconds, so cold-loading a large local model no longer errors out mid-load (from SDK v0.0.69)
• Empty Ollama responses are now retried instead of failing the task with "Model returned empty response" (from SDK v0.0.69)
• Migrated users whose stored Cline model id isn't in the catalog now fall back to the default model instead of sending an unknown model id on every request (from SDK v0.0.69)