spacehub

surshar / claude-code

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

Built by Anthropic · github.com/anthropics

claude-code / CHANGELOG.md
GitHub Actionschore: Update CHANGELOG.md da80366 · on Mar 6, 2026 History
2,101 lines · 129.3 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.70
4
5- Fixed API 400 errors when using `ANTHROPIC_BASE_URL` with a third-party gateway — tool search now correctly detects proxy endpoints and disables `tool_reference` blocks
6- Fixed `API Error: 400 This model does not support the effort parameter` when using custom Bedrock inference profiles or other model identifiers not matching standard Claude naming patterns
7- Fixed empty model responses immediately after `ToolSearch` — the server renders tool schemas with system-prompt-style tags at the prompt tail, which could confuse models into stopping early
8- Fixed prompt-cache bust when an MCP server with `instructions` connects after the first turn
9- Fixed Enter inserting a newline instead of submitting when typing over a slow SSH connection
10- Fixed clipboard corrupting non-ASCII text (CJK, emoji) on Windows/WSL by using PowerShell `Set-Clipboard`
11- Fixed extra VS Code windows opening at startup on Windows when running from the VS Code integrated terminal
12- Fixed voice mode failing on Windows native binary with "native audio module could not be loaded"
13- Fixed push-to-talk not activating on session start when `voiceEnabled: true` was set in settings
14- Fixed markdown links containing `#NNN` references incorrectly pointing to the current repository instead of the linked URL
15- Fixed repeated "Model updated to Opus 4.6" notification when a project's `.claude/settings.json` has a legacy Opus model string pinned
16- Fixed plugins showing as inaccurately installed in `/plugin`
17- Fixed plugins showing "not found in marketplace" errors on fresh startup by auto-refreshing after marketplace installation
18- Fixed `/security-review` command failing with `unknown option merge-base` on older git versions
19- Fixed `/color` command having no way to reset back to the default color — `/color default`, `/color gray`, `/color reset`, and `/color none` now restore the default
20- Fixed a performance regression in the `AskUserQuestion` preview dialog that re-ran markdown rendering on every keystroke in the notes input
21- Fixed feature flags read during early startup never refreshing their disk cache, causing stale values to persist across sessions
22- Fixed `permissions.defaultMode` settings values other than `acceptEdits` or `plan` being applied in Claude Code Remote environments — they are now ignored
23- Fixed skill listing being re-injected on every `--resume` (~600 tokens saved per resume)
24- Fixed teleport marker not rendering in VS Code teleported sessions
25- Improved error message when microphone captures silence to distinguish from "no speech detected"
26- Improved compaction to preserve images in the summarizer request, allowing prompt cache reuse for faster and cheaper compaction
27- Improved `/rename` to work while Claude is processing, instead of being silently queued
28- Reduced prompt input re-renders during turns by ~74%
29- Reduced startup memory by ~426KB for users without custom CA certificates
30- Reduced Remote Control `/poll` rate to once per 10 minutes while connected (was 1–2s), cutting server load ~300×. Reconnection is unaffected — transport loss immediately wakes fast polling.
31- [VSCode] Added spark icon in VS Code activity bar that lists all Claude Code sessions, with sessions opening as full editors
32- [VSCode] Added full markdown document view for plans in VS Code, with support for adding comments to provide feedback
33- [VSCode] Added native MCP server management dialog — use `/mcp` in the chat panel to enable/disable servers, reconnect, and manage OAuth authentication without switching to the terminal
34
35## 2.1.69
36
37- Added the `/claude-api` skill for building applications with the Claude API and Anthropic SDK
38- Added Ctrl+U on an empty bash prompt (`!`) to exit bash mode, matching `escape` and `backspace`
39- Added numeric keypad support for selecting options in Claude's interview questions (previously only the number row above QWERTY worked)
40- Added optional name argument to `/remote-control` and `claude remote-control` (`/remote-control My Project` or `--name "My Project"`) to set a custom session title visible in claude.ai/code
41- Added Voice STT support for 10 new languages (20 total) — Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, Norwegian
42- Added effort level display (e.g., "with low effort") to the logo and spinner, making it easier to see which effort setting is active
43- Added agent name display in terminal title when using `claude --agent`
44- Added `sandbox.enableWeakerNetworkIsolation` setting (macOS only) to allow Go programs like `gh`, `gcloud`, and `terraform` to verify TLS certificates when using a custom MITM proxy with `httpProxyPort`
45- Added `includeGitInstructions` setting (and `CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS` env var) to remove built-in commit and PR workflow instructions from Claude's system prompt
46- Added `/reload-plugins` command to activate pending plugin changes without restarting
47- Added a one-time startup prompt suggesting Claude Code Desktop on macOS and Windows (max 3 showings, dismissible)
48- Added `${CLAUDE_SKILL_DIR}` variable for skills to reference their own directory in SKILL.md content
49- Added `InstructionsLoaded` hook event that fires when CLAUDE.md or `.claude/rules/*.md` files are loaded into context
50- Added `agent_id` (for subagents) and `agent_type` (for subagents and `--agent`) to hook events
51- Added `worktree` field to status line hook commands with name, path, branch, and original repo directory when running in a `--worktree` session
52- Added `pluginTrustMessage` in managed settings to append organization-specific context to the plugin trust warning shown before installation
53- Added policy limit fetching (e.g., remote control restrictions) for Team plan OAuth users, not just Enterprise
54- Added `pathPattern` to `strictKnownMarketplaces` for regex-matching file/directory marketplace sources alongside `hostPattern` restrictions
55- Added plugin source type `git-subdir` to point to a subdirectory within a git repo
56- Added `oauth.authServerMetadataUrl` config option for MCP servers to specify a custom OAuth metadata discovery URL when standard discovery fails
57- Fixed a security issue where nested skill discovery could load skills from gitignored directories like `node_modules`
58- Fixed trust dialog silently enabling all `.mcp.json` servers on first run. You'll now see the per-server approval dialog as expected
59- Fixed `claude remote-control` crashing immediately on npm installs with "bad option: --sdk-url" (anthropics/claude-code#28334)
60- Fixed `--model claude-opus-4-0` and `--model claude-opus-4-1` resolving to deprecated Opus versions instead of current
61- Fixed macOS keychain corruption when using multiple OAuth MCP servers. Large OAuth metadata blobs could overflow the `security -i` stdin buffer, silently leaving stale credentials behind and causing repeated `/login` prompts.
62- Fixed `.credentials.json` losing `subscriptionType` (showing "Claude API" instead of "Claude Pro"/"Claude Max") when the profile endpoint transiently fails during token refresh (anthropics/claude-code#30185)
63- Fixed ghost dotfiles (`.bashrc`, `HEAD`, etc.) appearing as untracked files in the working directory after sandboxed Bash commands on Linux
64- Fixed Shift+Enter printing `[27;2;13~` instead of inserting a newline in Ghostty over SSH
65- Fixed stash (Ctrl+S) being cleared when submitting a message while Claude is working
66- Fixed ctrl+o (transcript toggle) freezing for many seconds in long sessions with lots of file edits
67- Fixed plan mode feedback input not supporting multi-line text entry (backslash+Enter and Shift+Enter now insert newlines)
68- Fixed cursor not moving down into blank lines at the top of the input box
69- Fixed `/stats` crash when transcript files contain entries with missing or malformed timestamps
70- Fixed a brief hang after a streaming error on long sessions (the transcript was being fully rewritten to drop one line; it is now truncated in place)
71- Fixed `--setting-sources user` not blocking dynamically discovered project skills
72- Fixed duplicate CLAUDE.md, slash commands, agents, and rules when running from a worktree nested inside its main repo (e.g. `claude -w`)
73- Fixed plugin Stop/SessionEnd/etc hooks not firing after any `/plugin` operation
74- Fixed plugin hooks being silently dropped when two plugins use the same `${CLAUDE_PLUGIN_ROOT}/...` command template
75- Fixed memory leak in long-running SDK/CCR sessions where conversation messages were retained unnecessarily
76- Fixed API 400 errors in forked agents (autocompact, summarization) when resuming sessions that were interrupted mid-tool-batch
77- Fixed "unexpected tool_use_id found in tool_result blocks" error when resuming conversations that start with an orphaned tool result
78- Fixed teammates accidentally spawning nested teammates via the Agent tool's `name` parameter
79- Fixed `CLAUDE_CODE_MAX_OUTPUT_TOKENS` being ignored during conversation compaction
80- Fixed `/compact` summary rendering as a user bubble in SDK consumers (Claude Code Remote web UI, VSCode extension)
81- Fixed voice space bar getting stuck after a failed voice activation (module loading race, cold GrowthBook)
82- Fixed worktree file copy on Windows
83- Fixed global `.claude` folder detection on Windows
84- Fixed symlink bypass where writing new files through a symlinked parent directory could escape the working directory in `acceptEdits` mode
85- Fixed sandbox prompting users to approve non-allowed domains when `allowManagedDomainsOnly` is enabled in managed settings — non-allowed domains are now blocked automatically with no bypass
86- Fixed interactive tools (e.g., `AskUserQuestion`) being silently auto-allowed when listed in a skill's allowed-tools, bypassing the permission prompt and running with empty answers
87- Fixed multi-GB memory spike when committing with large untracked binary files in the working tree
88- Fixed Escape not interrupting a running turn when the input box has draft text. Use Up arrow to pull queued messages back for editing, or Ctrl+U to clear the input line.
89- Fixed Android app crash when running local slash commands (`/voice`, `/cost`) in Remote Control sessions
90- Fixed a memory leak where old message array versions accumulated in React Compiler `memoCache` over long sessions
91- Fixed a memory leak where REPL render scopes accumulated over long sessions (~35MB over 1000 turns)
92- Fixed memory retention in in-process teammates where the parent's full conversation history was pinned for the teammate's lifetime, preventing GC after `/clear` or auto-compact
93- Fixed a memory leak in interactive mode where hook events could accumulate unboundedly during long sessions
94- Fixed hang when `--mcp-config` points to a corrupted file
95- Fixed slow startup when many skills/plugins are installed
96- Fixed `cd <outside-dir> && <cmd>` permission prompt to surface the chained command instead of only showing "Yes, allow reading from <dir>/"
97- Fixed conditional `.claude/rules/*.md` files (with `paths:` frontmatter) and nested CLAUDE.md files not loading in print mode (`claude -p`)
98- Fixed `/clear` not fully clearing all session caches, reducing memory retention in long sessions
99- Fixed terminal flicker caused by animated elements at the scrollback boundary
100- Fixed UI frame drops on macOS when using MCP servers with OAuth (regression from 2.1.x)
101- Fixed occasional frame stalls during typing caused by synchronous debug log flushes
102- Fixed `TeammateIdle` and `TaskCompleted` hooks to support `{"continue": false, "stopReason": "..."}` to stop the teammate, matching `Stop` hook behavior
103- Fixed `WorktreeCreate` and `WorktreeRemove` plugin hooks being silently ignored
104- Fixed skill descriptions with colons (e.g., "Triggers include: X, Y, Z") failing to load from SKILL.md frontmatter
105- Fixed project skills without a `description:` frontmatter field not appearing in Claude's available skills list
106- Fixed `/context` showing identical token counts for all MCP tools from a server
107- Fixed literal `nul` file creation on Windows when the model uses CMD-style `2>nul` redirection in Git Bash
108- Fixed extra blank lines appearing below each tool call in the expanded subagent transcript view (Ctrl+O)
109- Fixed Tab/arrow keys not cycling Settings tabs when `/config` search box is focused but empty
110- Fixed service key OAuth sessions (CCR containers) spamming `[ERROR]` logs with 403s from profile-scoped endpoints
111- Fixed inconsistent color for "Remote Control active" status indicator
112- Fixed Voice waveform cursor covering the first suffix letter when dictating mid-input
113- Fixed Voice input showing all 5 spaces during warmup instead of capping at ~2 (aligning with the "keep holding…" hint)
114- Improved spinner performance by isolating the 50ms animation loop from the surrounding shell, reducing render and CPU overhead during turns
115- Improved UI rendering performance in native binaries with React Compiler
116- Improved `--worktree` startup by eliminating a git subprocess on the startup path
117- Improved macOS startup by eliminating redundant settings-file reloads when managed settings resolve
118- Improved macOS startup for Claude.ai enterprise/team users by skipping an unnecessary keychain lookup
119- Improved MCP `-p` startup by pipelining claude.ai config fetch with local connections and using a concurrency pool instead of sequential batching
120- Improved voice startup by removing imperceptible warmup pulse animations that were causing re-render stutter
121- Improved MCP binary content handling: tools returning PDFs, Office documents, or audio now save decoded bytes to disk with the correct file extension instead of dumping raw base64 into the conversation context. WebFetch also saves binary responses alongside its summary.
122- Improved memory usage in long sessions by stabilizing `onSubmit` across message updates
123- Improved LSP tool rendering and memory context building to no longer read entire files
124- Improved session upload and memory sync to avoid reading large files into memory before size/binary checks
125- Improved file operation performance by avoiding reading file contents for existence checks (6 sites)
126- Improved documentation to clarify that `--append-system-prompt-file` and `--system-prompt-file` work in interactive mode (the docs previously said print mode only)
127- Reduced baseline memory by ~16MB by deferring Yoga WASM preloading
128- Reduced memory footprint for SDK and CCR sessions using stream-json output
129- Reduced memory usage when resuming large sessions (including compacted history)
130- Reduced token usage on multi-agent tasks with more concise subagent final reports
131- Changed Sonnet 4.5 users on Pro/Max/Team Premium to be automatically migrated to Sonnet 4.6
132- Changed the `/resume` picker to show your most recent prompt instead of the first one. This also resolves some titles appearing as `(session)`.
133- Changed claude.ai MCP connector failures to show a notification instead of silently disappearing from the tool list
134- Changed example command suggestions to be generated deterministically instead of calling Haiku
135- Changed resuming after compaction to no longer produce a preamble recap before continuing
136- [SDK] Changed task creation to no longer require the `activeForm` field — the spinner falls back to the task subject
137- [VSCode] Added compaction display as a collapsible "Compacted chat" card with the summary inside
138- [VSCode] The permission mode picker now respects `permissions.disableBypassPermissionsMode` from your effective Claude Code settings (including managed/policy settings) — when set to `disable`, bypass permissions mode is hidden from the picker
139- [VSCode] Fixed RTL text (Arabic, Hebrew, Persian) rendering reversed in the chat panel (regression in v2.1.63)
140
141## 2.1.68
142
143- Opus 4.6 now defaults to medium effort for Max and Team subscribers. Medium effort works well for most tasks — it's the sweet spot between speed and thoroughness. You can change this anytime with `/model`
144- Re-introduced the "ultrathink" keyword to enable high effort for the next turn
145- Removed Opus 4 and 4.1 from Claude Code on the first-party API — users with these models pinned are automatically moved to Opus 4.6
146
147## 2.1.66
148
149- Reduced spurious error logging
150
151## 2.1.63
152
153- Added `/simplify` and `/batch` bundled slash commands
154- Fixed local slash command output like /cost appearing as user-sent messages instead of system messages in the UI
155- Project configs & auto memory now shared across git worktrees of the same repository
156- Added `ENABLE_CLAUDEAI_MCP_SERVERS=false` env var to opt out from making claude.ai MCP servers available
157- Improved `/model` command to show the currently active model in the slash command menu
158- Added HTTP hooks, which can POST JSON to a URL and receive JSON instead of running a shell command
159- Fixed listener leak in bridge polling loop
160- Fixed listener leak in MCP OAuth flow cleanup
161- Added manual URL paste fallback during MCP OAuth authentication. If the automatic localhost redirect doesn't work, you can paste the callback URL to complete authentication.
162- Fixed memory leak when navigating hooks configuration menu
163- Fixed listener leak in interactive permission handler during auto-approvals
164- Fixed file count cache ignoring glob ignore patterns
165- Fixed memory leak in bash command prefix cache
166- Fixed MCP tool/resource cache leak on server reconnect
167- Fixed IDE host IP detection cache incorrectly sharing results across ports
168- Fixed WebSocket listener leak on transport reconnect
169- Fixed memory leak in git root detection cache that could cause unbounded growth in long-running sessions
170- Fixed memory leak in JSON parsing cache that grew unbounded over long sessions
171- VSCode: Fixed remote sessions not appearing in conversation history
172- Fixed a race condition in the REPL bridge where new messages could arrive at the server interleaved with historical messages during the initial connection flush, causing message ordering issues.
173- Fixed memory leak where long-running teammates retained all messages in AppState even after conversation compaction
174- Fixed a memory leak where MCP server fetch caches were not cleared on disconnect, causing growing memory usage with servers that reconnect frequently
175- Improved memory usage in long sessions with subagents by stripping heavy progress message payloads during context compaction
176- Added "Always copy full response" option to the `/copy` picker. When selected, future `/copy` commands will skip the code block picker and copy the full response directly.
177- VSCode: Added session rename and remove actions to the sessions list
178- Fixed `/clear` not resetting cached skills, which could cause stale skill content to persist in the new conversation
179
180## 2.1.62
181
182- Fixed prompt suggestion cache regression that reduced cache hit rates
183
184## 2.1.61
185
186- Fixed concurrent writes corrupting config file on Windows
187
188## 2.1.59
189
190- Claude automatically saves useful context to auto-memory. Manage with /memory
191- Added `/copy` command to show an interactive picker when code blocks are present, allowing selection of individual code blocks or the full response.
192- Improved "always allow" prefix suggestions for compound bash commands (e.g. `cd /tmp && git fetch && git push`) to compute smarter per-subcommand prefixes instead of treating the whole command as one
193- Improved ordering of short task lists
194- Improved memory usage in multi-agent sessions by releasing completed subagent task state
195- Fixed MCP OAuth token refresh race condition when running multiple Claude Code instances simultaneously
196- Fixed shell commands not showing a clear error message when the working directory has been deleted
197- Fixed config file corruption that could wipe authentication when multiple Claude Code instances ran simultaneously
198
199## 2.1.58
200
201- Expand Remote Control to more users
202
203## 2.1.56
204
205- VS Code: Fixed another cause of "command 'claude-vscode.editor.openLast' not found" crashes
206
207## 2.1.55
208
209- Fixed BashTool failing on Windows with EINVAL error
210
211## 2.1.53
212
213- Fixed a UI flicker where user input would briefly disappear after submission before the message rendered
214- Fixed bulk agent kill (ctrl+f) to send a single aggregate notification instead of one per agent, and to properly clear the command queue
215- Fixed graceful shutdown sometimes leaving stale sessions when using Remote Control by parallelizing teardown network calls
216- Fixed `--worktree` sometimes being ignored on first launch
217- Fixed a panic ("switch on corrupted value") on Windows
218- Fixed a crash that could occur when spawning many processes on Windows
219- Fixed a crash in the WebAssembly interpreter on Linux x64 & Windows x64
220- Fixed a crash that sometimes occurred after 2 minutes on Windows ARM64
221
222## 2.1.52
223
224- VS Code: Fixed extension crash on Windows ("command 'claude-vscode.editor.openLast' not found")
225
226## 2.1.51
227
228- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
229- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
230- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
231- BashTool now skips login shell (`-l` flag) by default when a shell snapshot is available, improving command execution performance. Previously this required setting `CLAUDE_BASH_NO_LOGIN=true`.
232- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
233- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
234- Fixed a bug where duplicate `control_response` messages (e.g. from WebSocket reconnects) could cause API 400 errors by pushing duplicate assistant messages into the conversation.
235- Added `CLAUDE_CODE_ACCOUNT_UUID`, `CLAUDE_CODE_USER_EMAIL`, and `CLAUDE_CODE_ORGANIZATION_UUID` environment variables for SDK callers to provide account info synchronously, eliminating a race condition where early telemetry events lacked account metadata.
236- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
237- The `/model` picker now shows human-readable labels (e.g., "Sonnet 4.5") instead of raw model IDs for pinned model versions, with an upgrade hint when a newer version is available.
238- Managed settings can now be set via macOS plist or Windows Registry. Learn more at https://code.claude.com/docs/en/settings#settings-files
239
240## 2.1.50
241
242- Added support for `startupTimeout` configuration for LSP servers
243- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
244- Fixed a bug where resumed sessions could be invisible when the working directory involved symlinks, because the session storage path was resolved at different times during startup. Also fixed session data loss on SSH disconnect by flushing session data before hooks and analytics in the graceful shutdown sequence.
245- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
246- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
247- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
248- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
249- Fixed memory leak where completed task state objects were never removed from AppState
250- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
251- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
252- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
253- Improved memory usage during long sessions by clearing internal caches after compaction
254- Added `claude agents` CLI command to list all configured agents
255- Improved memory usage during long sessions by clearing large tool results after they have been processed
256- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
257- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
258- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
259- Fixed prompt suggestion cache regression that reduced cache hit rates
260- Fixed unbounded memory growth in long sessions by capping file history snapshots
261- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
262- Opus 4.6 (fast mode) now includes the full 1M context window
263- VSCode: Added `/extra-usage` command support in VS Code sessions
264- Fixed memory leak where TaskOutput retained recent lines after cleanup
265- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
266- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
267
268## 2.1.49
269
270- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
271- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
272- Subagents support `isolation: "worktree"` for working in a temporary git worktree
273- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
274- Agent definitions support `background: true` to always run as a background task
275- Plugins can ship `settings.json` for default configuration
276- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
277- Fixed Ctrl+C and ESC being silently ignored when background agents are running and the main thread is idle. Pressing twice within 3 seconds now kills all background agents.
278- Fixed prompt suggestion cache regression that reduced cache hit rates.
279- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
280- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
281- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
282- Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in /model.
283- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
284- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
285- Fixed potential rendering issues caused by stale yoga layout references
286- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
287- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
288- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
289- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
290- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
291- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
292- Improved startup performance by reducing HTTP calls for analytics token counting
293- Improved startup performance by batching MCP tool token counting into a single API call
294- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
295- Fixed `--resume` session picker showing raw XML tags for sessions that start with commands like `/clear`. Now correctly falls through to the session ID fallback.
296- Improved permission prompts for path safety and working directory blocks to show the reason for the restriction instead of a bare prompt with no context
297
298## 2.1.47
299
300- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
301- Fixed Windows terminal rendering bugs caused by `os.EOL` (`\r\n`) in display code — line counts now show correct values instead of always showing 1 on Windows.
302- Improved VS Code plan preview: auto-updates as Claude iterates, enables commenting only when the plan is ready for review, and keeps the preview open when rejecting so Claude can revise.
303- Fixed a bug where bold and colored text in markdown output could shift to the wrong characters on Windows due to `\r\n` line endings.
304- Fixed compaction failing when conversation contains many PDF documents by stripping document blocks alongside images before sending to the compaction API (anthropics/claude-code#26188)
305- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
306- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
307- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
308- Improved performance of `@` file mentions - file suggestions now appear faster by pre-warming the index on startup and using session-based caching with background refresh.
309- Improved memory usage by trimming agent task message history after tasks complete
310- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
311- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
312- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
313- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
314- Search patterns in collapsed tool results are now displayed in quotes for clarity
315- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
316- Use `ctrl+f` to kill all background agents instead of double-pressing ESC. Background agents now continue running when you press ESC to cancel the main thread, giving you more control over agent lifecycle.
317- Fixed API 400 errors ("thinking blocks cannot be modified") that occurred in sessions with concurrent agents, caused by interleaved streaming content blocks preventing proper message merging.
318- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
319- Fixed an issue where a single file write/edit error would abort all other parallel file write/edit operations. Independent file mutations now complete even when a sibling fails.
320- Added `last_assistant_message` field to Stop and SubagentStop hook inputs, providing the final assistant response text so hooks can access it without parsing transcript files.
321- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
322- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
323- Fixed an issue where bash commands with backslash-newline continuation lines (e.g., long commands split across multiple lines with `\`) would produce spurious empty arguments, potentially breaking command execution.
324- Fixed built-in slash commands (`/help`, `/model`, `/compact`, etc.) being hidden from the autocomplete dropdown when many user skills are installed (anthropics/claude-code#22020)
325- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
326- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
327- Fixed crash when a skill's `name` or `description` in SKILL.md frontmatter is a bare number (e.g., `name: 3000`) — the value is now properly coerced to a string (anthropics/claude-code#25837)
328- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
329- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
330- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
331- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
332- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
333- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
334- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
335- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
336- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
337- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
338- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
339- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
340- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
341- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
342- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
343- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
344- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
345- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
346- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
347- Fixed React crash (error #31) when a skill's `argument-hint` in SKILL.md frontmatter uses YAML sequence syntax (e.g., `[topic: foo | bar]`) — the value is now properly coerced to a string (anthropics/claude-code#25826)
348- Fixed crash when using `/fork` on sessions that used web search — null entries in search results from transcript deserialization are now handled gracefully (anthropics/claude-code#25811)
349- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
350- Fixed custom agents and skills not being discovered when running from a git worktree — project-level `.claude/agents/` and `.claude/skills/` from the main repository are now included (anthropics/claude-code#25816)
351- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
352- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
353- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
354- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
355- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
356- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
357- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
358- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
359- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
360- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
361- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
362- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
363- Fixed background tasks failing in git worktrees due to remote URL resolution reading from worktree-specific gitdir instead of the main repository config (anthropics/claude-code#26065)
364- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
365- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
366- Fixed Edit tool silently corrupting Unicode curly quotes (\u201c\u201d \u2018\u2019) by replacing them with straight quotes when making edits (anthropics/claude-code#26141)
367- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
368
369## 2.1.46
370
371- Fixed orphaned CC processes after terminal disconnect on macOS
372- Added support for using claude.ai MCP connectors in Claude Code
373
374## 2.1.45
375
376- Added support for Claude Sonnet 4.6
377- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
378- Added `spinnerTipsOverride` setting to customize spinner tips — configure `tips` with an array of custom tip strings, and optionally set `excludeDefault: true` to show only your custom tips instead of the built-in ones
379- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
380- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
381- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
382- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
383- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
384- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
385- Fixed excessive `.claude.json.backup` files accumulating on every startup
386- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
387- Improved startup performance by removing eager loading of session history for stats caching
388- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
389- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
390- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
391
392## 2.1.44
393
394- Fixed ENAMETOOLONG errors for deeply-nested directory paths
395- Fixed auth refresh errors
396
397## 2.1.43
398
399- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
400- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
401- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
402
403## 2.1.42
404
405- Improved startup performance by deferring Zod schema construction
406- Improved prompt cache hit rates by moving date out of system prompt
407- Added one-time Opus 4.6 effort callout for eligible users
408- Fixed /resume showing interrupt messages as session titles
409- Fixed image dimension limit errors to suggest /compact
410
411## 2.1.41
412
413- Added guard against launching Claude Code inside another Claude Code session
414- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
415- Fixed a crash when MCP tools return image content during streaming
416- Fixed /resume session previews showing raw XML tags instead of readable command names
417- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
418- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
419- Fixed hook blocking errors (exit code 2) not showing stderr to the user
420- Added `speed` attribute to OTel events and trace spans for fast mode visibility
421- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
422- Added Windows ARM64 (win32-arm64) native binary support
423- Improved `/rename` to auto-generate session name from conversation context when called without arguments
424- Improved narrow terminal layout for prompt footer
425- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
426- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
427- Fixed background task notifications not being delivered in streaming Agent SDK mode
428- Fixed cursor jumping to end on each keystroke in classifier rule input
429- Fixed markdown link display text being dropped for raw URL
430- Fixed auto-compact failure error notifications being shown to users
431- Fixed permission wait time being included in subagent elapsed time display
432- Fixed proactive ticks firing while in plan mode
433- Fixed clear stale permission rules when settings change on disk
434- Fixed hook blocking errors showing stderr content in UI
435
436## 2.1.39
437
438- Improved terminal rendering performance
439- Fixed fatal errors being swallowed instead of displayed
440- Fixed process hanging after session close
441- Fixed character loss at terminal screen boundary
442- Fixed blank lines in verbose transcript view
443
444## 2.1.38
445
446- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
447- Fixed Tab key queueing slash commands instead of autocompleting
448- Fixed bash permission matching for commands using environment variable wrappers
449- Fixed text between tool uses disappearing when not using streaming
450- Fixed duplicate sessions when resuming in VS Code extension
451- Improved heredoc delimiter parsing to prevent command smuggling
452- Blocked writes to `.claude/skills` directory in sandbox mode
453
454## 2.1.37
455
456- Fixed an issue where /fast was not immediately available after enabling /extra-usage
457
458## 2.1.36
459
460- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
461
462## 2.1.34
463
464- Fixed a crash when agent teams setting changed between renders
465- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
466
467## 2.1.33
468
469- Fixed agent teammate sessions in tmux to send and receive messages
470- Fixed warnings about agent teams not being available on your current plan
471- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
472- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
473- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
474- Added plugin name to skill descriptions and `/skills` menu for better discoverability
475- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
476- Fixed an API error that could occur when aborting mid-stream, where whitespace text combined with a thinking block would bypass normalization and produce an invalid request
477- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
478- Fixed an issue where proxy settings configured via `settings.json` environment variables were not applied to WebFetch and other HTTP requests on the Node.js build
479- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
480- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
481- Errors from invalid managed settings are now surfaced
482- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
483- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
484- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
485
486## 2.1.32
487
488- Claude Opus 4.6 is now available!
489- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
490- Claude now automatically records and recalls memories as it works
491- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
492- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
493- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
494- Updated --resume to re-use --agent value specified in previous conversation by default.
495- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
496- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
497- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
498- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
499- VSCode: Added spinner when loading past conversations list
500
501## 2.1.31
502
503- Added session resume hint on exit, showing how to continue your conversation later
504- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
505- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
506- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
507- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
508- Fixed `temperatureOverride` being silently ignored in the streaming API path, causing all streaming requests to use the default temperature (1) regardless of the configured override
509- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
510- Improved system prompts to more clearly guide the model toward using dedicated tools (Read, Edit, Glob, Grep) instead of bash equivalents (`cat`, `sed`, `grep`, `find`), reducing unnecessary bash command usage
511- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
512- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
513- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
514
515## 2.1.30
516
517- Added `pages` parameter to the Read tool for PDFs, allowing specific page ranges to be read (e.g., `pages: "1-5"`). Large PDFs (>10 pages) now return a lightweight reference when `@` mentioned instead of being inlined into context.
518- Added pre-configured OAuth client credentials for MCP servers that don't support Dynamic Client Registration (e.g., Slack). Use `--client-id` and `--client-secret` with `claude mcp add`.
519- Added `/debug` for Claude to help troubleshoot the current session
520- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
521- Added token count, tool uses, and duration metrics to Task tool results
522- Added reduced motion mode to the config
523- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
524- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
525- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
526- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
527- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
528- Fixed permission dialogs stealing focus while actively typing
529- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
530- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
531- Improved memory usage for `--resume` (68% reduction for users with many sessions) by replacing the session index with lightweight stat-based loading and progressive enrichment
532- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
533- Changed `/model` to execute immediately instead of being queued
534- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
535- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
536
537## 2.1.29
538
539- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
540
541## 2.1.27
542
543- Added tool call failures and denials to debug logs
544- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
545- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
546- Sessions are now automatically linked to PRs when created via `gh pr create`
547- Fixed /context command not displaying colored output
548- Fixed status bar duplicating background task indicator when PR status was shown
549- Windows: Fixed bash command execution failing for users with `.bashrc` files
550- Windows: Fixed console windows flashing when spawning child processes
551- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
552
553## 2.1.25
554
555- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
556
557## 2.1.23
558
559- Added customizable spinner verbs setting (`spinnerVerbs`)
560- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
561- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
562- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
563- Fixed pending async hooks not being cancelled when headless streaming sessions ended
564- Fixed tab completion not updating the input field when accepting a suggestion
565- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
566- Improved terminal rendering performance with optimized screen data layout
567- Changed Bash commands to show timeout duration alongside elapsed time
568- Changed merged pull requests to show a purple status indicator in the prompt footer
569- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
570
571## 2.1.22
572
573- Fixed structured outputs for non-interactive (-p) mode
574
575## 2.1.21
576
577- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
578- Fixed shell completion cache files being truncated on exit
579- Fixed API errors when resuming sessions that were interrupted during tool execution
580- Fixed auto-compact triggering too early on models with large output token limits
581- Fixed task IDs potentially being reused after deletion
582- Fixed file search not working in VS Code extension on Windows
583- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
584- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
585- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
586- [VSCode] Fixed message action buttons having incorrect background colors
587
588## 2.1.20
589
590- Added arrow key history navigation in vim normal mode when cursor cannot move further
591- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
592- Added PR review status indicator to the prompt footer, showing the current branch's PR state (approved, changes requested, pending, or draft) as a colored dot with a clickable link
593- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
594- Added ability to delete tasks via the `TaskUpdate` tool
595- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
596- Fixed agents sometimes ignoring user messages sent while actively working on a task
597- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
598- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
599- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
600- Fixed draft prompt being lost when pressing UP arrow to navigate command history
601- Fixed ghost text flickering when typing slash commands mid-input
602- Fixed marketplace source removal not properly deleting settings
603- Fixed duplicate output in some commands like `/context`
604- Fixed task list sometimes showing outside the main conversation view
605- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
606- Fixed crashes when cancelling tool use
607- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
608- Improved thinking status text with a subtle shimmer animation
609- Improved task list to dynamically adjust visible items based on terminal height
610- Improved fork conversation hint to show how to resume the original session
611- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
612- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
613- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
614- Changed the `/copy` command to be available to all users
615- Changed background agents to prompt for tool permissions before launching
616- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
617- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
618
619## 2.1.19
620
621- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
622- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
623- Fixed crashes on processors without AVX instruction support
624- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
625- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
626- Fixed resuming sessions by custom title not working when run from a different directory
627- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
628- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
629- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
630- Fixed file write preview omitting empty lines
631- Changed skills without additional permissions or hooks to be allowed without requiring approval
632- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
633- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
634- [VSCode] Enabled session forking and rewind functionality for all users
635
636## 2.1.18
637
638- Added customizable keyboard shortcuts. Configure keybindings per context, create chord sequences, and personalize your workflow. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
639
640## 2.1.17
641
642- Fixed crashes on processors without AVX instruction support
643
644## 2.1.16
645
646- Added new task management system, including new capabilities like dependency tracking
647- [VSCode] Added native plugin management support
648- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
649- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
650- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
651- Fixed session titles on the resume screen not respecting the user's language setting
652- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
653
654## 2.1.15
655
656- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
657- Improved UI rendering performance with React Compiler
658- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
659- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
660
661## 2.1.14
662
663- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
664- Added search to installed plugins list - type to filter by name or description
665- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
666- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
667- Fixed memory issues that could cause crashes when running parallel subagents
668- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
669- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
670- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
671- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
672- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
673- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
674- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
675- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
676- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
677- Improved backspace to delete pasted text as a single token instead of one character at a time
678- [VSCode] Added `/usage` command to display current plan usage
679
680## 2.1.12
681
682- Fixed message rendering bug
683
684## 2.1.11
685
686- Fixed excessive MCP connection requests for HTTP/SSE transports
687
688## 2.1.10
689
690- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
691- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
692- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
693- Improved startup to capture keystrokes typed before the REPL is fully ready
694- Improved file suggestions to show as removable attachments instead of inserting text when accepted
695- [VSCode] Added install count display to plugin listings
696- [VSCode] Added trust warning when installing plugins
697
698## 2.1.9
699
700- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
701- Added `plansDirectory` setting to customize where plan files are stored
702- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
703- Added session URL attribution to commits and PRs created from web sessions
704- Added support for `PreToolUse` hooks to return `additionalContext` to the model
705- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
706- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
707- Fixed MCP server reconnection hanging when cached connection promise never resolves
708- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
709
710## 2.1.7
711
712- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
713- Added ability to provide feedback when accepting permission prompts
714- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
715- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
716- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
717- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
718- Fixed context window blocking limit being calculated using the full context window instead of the effective context window (which reserves space for max output tokens)
719- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
720- Fixed terminal title animation jitter by using fixed-width braille characters
721- Fixed plugins with git submodules not being fully initialized when installed
722- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
723- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
724- Enabled MCP tool search auto mode by default for all users. When MCP tool descriptions exceed 10% of the context window, they are automatically deferred and discovered via the MCPSearch tool instead of being loaded upfront. This reduces context usage for users with many MCP tools configured. Users can disable this by adding `MCPSearch` to `disallowedTools` in their settings.
725- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
726- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
727
728## 2.1.6
729
730- Added search functionality to `/config` command for quickly filtering settings
731- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
732- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
733- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
734- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
735- Added an error display when the editor fails during Ctrl+G
736- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
737- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
738- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
739- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
740- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
741- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
742- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
743- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
744- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
745- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
746- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
747- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
748- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
749- Improved @ autocomplete with icons for different suggestion types and single-line formatting
750- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
751- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
752- Changed terminal title to "Claude Code" on startup for better window identification
753- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
754- [VSCode] Fixed usage indicator not updating after manual compact
755
756## 2.1.5
757
758- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
759
760## 2.1.4
761
762- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
763- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
764
765## 2.1.3
766
767- Merged slash commands and skills, simplifying the mental model with no change in behavior
768- Added release channel (`stable` or `latest`) toggle to `/config`
769- Added detection and warnings for unreachable permission rules, with warnings in `/doctor` and after saving rules that include the source of each rule and actionable fix guidance
770- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
771- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
772- Fixed mismatch between background task count in status bar and items shown in tasks dialog
773- Fixed sub-agents using the wrong model during conversation compaction
774- Fixed web search in sub-agents using incorrect model
775- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
776- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
777- Improved slash command suggestion readability by truncating long descriptions to 2 lines
778- Changed tool hook execution timeout from 60 seconds to 10 minutes
779- [VSCode] Added clickable destination selector for permission requests, allowing you to choose where settings are saved (this project, all projects, shared with team, or session only)
780
781## 2.1.2
782
783- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
784- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
785- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
786- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
787- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
788- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
789- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
790- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
791- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
792- Fixed updates incorrectly claiming another installation is in progress
793- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
794- Fixed remote session URL and teleport being broken when using `/tasks` command
795- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
796- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
797- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
798- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
799- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
800- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
801- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
802- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
803- [SDK] Changed minimum zod peer dependency to ^4.0.0
804- [VSCode] Fixed usage display not updating after manual compact
805
806## 2.1.0
807
808- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
809- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
810- Added support for `agent` field in skills to specify agent type for execution
811- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
812- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
813- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
814- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
815- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
816- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
817- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
818- Fixed Esc key with queued prompts to only move them to input without canceling the running task
819- Reduced permission prompts for complex bash commands
820- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
821- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
822- Fixed plugin path resolution for file-based marketplace sources
823- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
824- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
825- Fixed Claude in Chrome support for WSL environments
826- Fixed Windows native installer silently failing when executable creation fails
827- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
828- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
829- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
830- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
831- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
832- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
833- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
834- Added hooks support for skill and slash command frontmatter
835- Added new Vim motions: `;` and `,` to repeat f/F/t/T motions, `y` operator for yank with `yy`/`Y`, `p`/`P` for paste, text objects (`iw`, `aw`, `iW`, `aW`, `i"`, `a"`, `i'`, `a'`, `i(`, `a(`, `i[`, `a[`, `i{`, `a{`), `>>` and `<<` for indent/dedent, and `J` to join lines
836- Added `/plan` command shortcut to enable plan mode directly from the prompt
837- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
838- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
839- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
840- Added support for `once: true` config for hooks
841- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
842- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
843- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
844- Added left/right arrow key navigation for cycling through tabs in dialogs
845- Added real-time thinking block display in Ctrl+O transcript mode
846- Added filepath to full output in background bash task details dialog
847- Added Skills as a separate category in the context visualization
848- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
849- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
850- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
851- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
852- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
853- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
854- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
855- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
856- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
857- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
858- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
859- Fixed commands with `$()` command substitution failing with parse errors
860- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
861- Fixed bash command prefix extraction to correctly identify subcommands after global options (e.g., `git -C /path log` now correctly matches `Bash(git log:*)` rules)
862- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
863- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
864- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
865- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
866- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
867- Fixed subagents sometimes not inheriting the parent's model by default
868- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
869- Fixed duplicate Bash commands appearing in permission request option labels
870- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
871- Fixed background task completion notifications to appear proactively with bullet point
872- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
873- Fixed cursor disappearing after dismissing permission dialogs
874- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
875- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
876- Fixed images being silently dropped when queueing messages while backgrounding a task
877- Fixed large pasted images failing with "Image was too large" error
878- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
879- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
880- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
881- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
882- Fixed spinner token counter to properly accumulate tokens from subagents during execution
883- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
884- Fixed race condition where LSP tool could return "no server available" during startup
885- Fixed feedback submission hanging indefinitely when network requests timeout
886- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
887- Fixed hook success message showing trailing colon when hook has no output
888- Multiple optimizations to improve startup performance
889- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
890- Improved performance when reading Jupyter notebooks with many cells
891- Improved reliability for piped input like `cat refactor.md | claude`
892- Improved reliability for AskQuestion tool
893- Improved sed in-place edit commands to render as file edits with diff preview
894- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
895- Improved compaction reliability
896- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
897- Improved skills to show progress while executing, displaying tool uses as they happen
898- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
899- Improved skill suggestions to prioritize recently and frequently used skills
900- Improved spinner feedback when waiting for the first response token
901- Improved token count display in spinner to include tokens from background agents
902- Improved incremental output for async agents to give the main thread more control and visibility
903- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
904- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
905- Improved macOS screenshot paste reliability with TIFF format support
906- Improved `/stats` output
907- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
908- Changed "Interrupted" message color from red to grey for a less alarming appearance
909- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
910- Removed underline styling from image reference links
911- [SDK] Changed minimum zod peer dependency to ^4.0.0
912- [VSCode] Added currently selected model name to the context menu
913- [VSCode] Added descriptive labels on auto-accept permission button (e.g., "Yes, allow npm for this project" instead of "Yes, and don't ask again")
914- [VSCode] Fixed paragraph breaks not rendering in markdown content
915- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
916- [Windows] Fixed issue with improper rendering
917
918## 2.0.76
919
920- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
921
922## 2.0.75
923
924- Minor bugfixes
925
926## 2.0.74
927
928- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
929- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
930- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
931- Added syntax highlighting info to theme picker
932- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
933- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
934- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
935- Fixed a potential crash when syntax highlighting isn't initialized correctly
936- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
937- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
938- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
939- [Windows] Fixed issue with improper rendering
940- [VSCode] Added gift tag pictogram for year-end promotion message
941
942## 2.0.73
943
944- Added clickable `[Image #N]` links that open attached images in the default viewer
945- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
946- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
947- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
948- Fixed slow input history cycling and race condition that could overwrite text after message submission
949- Improved `/theme` command to open theme picker directly
950- Improved theme picker UI
951- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
952- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
953
954## 2.0.72
955
956- Added Claude in Chrome (Beta) feature that works with the Chrome extension (https://claude.ai/chrome) to let you control your browser directly from Claude Code
957- Reduced terminal flickering
958- Added scannable QR code to mobile app tip for quick app downloads
959- Added loading indicator when resuming conversations for better feedback
960- Fixed `/context` command not respecting custom system prompts in non-interactive mode
961- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
962- Improved @ mention file suggestion speed (~3× faster in git repositories)
963- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
964- Improved settings validation errors to be more prominent
965- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
966
967## 2.0.71
968
969- Added /config toggle to enable/disable prompt suggestions
970- Added `/settings` as an alias for the `/config` command
971- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
972- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
973- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
974- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
975- New syntax highlighting engine for native build
976
977## 2.0.70
978
979- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
980- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
981- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
982- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
983- Fixed input being cleared when processing queued commands while the user was typing
984- Fixed prompt suggestions replacing typed input when pressing Tab
985- Fixed diff view not updating when terminal is resized
986- Improved memory usage by 3x for large conversations
987- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
988- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
989- Fix thinking mode toggle in /config not persisting correctly
990- Improve UI for file creation permission dialog
991
992## 2.0.69
993
994- Minor bugfixes
995
996## 2.0.68
997
998- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
999- Fixed a bug where disallowed MCP tools were visible to the model
1000- Fixed an issue where steering messages could be lost while a subagent is working
1001- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
1002- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
1003- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
1004
1005## 2.0.67
1006
1007- Thinking mode is now enabled by default for Opus 4.5
1008- Thinking mode configuration has moved to /config
1009- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
1010- Show reason why autoupdater is disabled in `/doctor`
1011- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
1012- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
1013- Fixed scroll position resetting after deleting a permission rule in `/permissions`
1014- Fixed word deletion (opt+delete) and word navigation (opt+arrow) not working correctly with non-Latin text such as Cyrillic, Greek, Arabic, Hebrew, Thai, and Chinese
1015- Fixed `claude install --force` not bypassing stale lock files
1016- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
1017- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
1018
1019## 2.0.65
1020
1021- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
1022- Added context window information to status line input
1023- Added `fileSuggestion` setting for custom `@` file search commands
1024- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
1025- Fixed prompt not being saved to history when aborting a query with Escape
1026- Fixed Read tool image handling to identify format from bytes instead of file extension
1027
1028## 2.0.64
1029
1030- Made auto-compacting instant
1031- Agents and bash commands can run asynchronously and send messages to wake up the main agent
1032- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
1033- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
1034- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
1035- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
1036- Fixed auto-loading .env when using native installer
1037- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
1038- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
1039- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
1040- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
1041- Bedrock: Improve efficiency of token counting
1042- Bedrock: Add support for `aws login` AWS Management Console credentials
1043- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
1044
1045## 2.0.62
1046
1047- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
1048- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
1049- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
1050- Fixed slash command selection not working when multiple commands share the same name
1051- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
1052- Fixed running versions getting removed because lock file incorrectly going stale
1053- Fixed IDE diff tab not closing when rejecting file changes
1054
1055## 2.0.61
1056
1057- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
1058
1059## 2.0.60
1060
1061- Added background agent support. Agents run in the background while you work
1062- Added --disable-slash-commands CLI flag to disable all slash commands
1063- Added model name to "Co-Authored-By" commit messages
1064- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
1065- Updated Fetch to skip summarization for pre-approved websites
1066- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
1067
1068## 2.0.59
1069
1070- Added --agent CLI flag to override the agent setting for the current session
1071- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
1072- VS Code: Fixed .claude.json config file being read from incorrect location
1073
1074## 2.0.58
1075
1076- Pro users now have access to Opus 4.5 as part of their subscription!
1077- Fixed timer duration showing "11m 60s" instead of "12m 0s"
1078- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
1079
1080## 2.0.57
1081
1082- Added feedback input when rejecting plans, allowing users to tell Claude what to change
1083- VSCode: Added streaming message support for real-time response display
1084
1085## 2.0.56
1086
1087- Added setting to enable/disable terminal progress bar (OSC 9;4)
1088- VSCode Extension: Added support for VS Code's secondary sidebar (VS Code 1.97+), allowing Claude Code to be displayed in the right sidebar while keeping the file explorer on the left. Requires setting sidebar as Preferred Location in the config.
1089
1090## 2.0.55
1091
1092- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
1093- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
1094- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
1095- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
1096
1097## 2.0.54
1098
1099- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
1100- Fix issue with excessive iTerm notifications
1101
1102## 2.0.52
1103
1104- Fixed duplicate message display when starting Claude with a command line argument
1105- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
1106- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
1107- Permit some uses of `$!` in bash commands
1108
1109## 2.0.51
1110
1111- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
1112- Introducing Claude Code for Desktop: https://claude.com/download
1113- To give you room to try out our new model, we've updated usage limits for Claude Code users. See the Claude Opus 4.5 blog for full details
1114- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
1115- Plan Mode now builds more precise plans and executes more thoroughly
1116- Usage limit notifications now easier to understand
1117- Switched `/usage` back to "% used"
1118- Fixed handling of thinking errors
1119- Fixed performance regression
1120
1121## 2.0.50
1122
1123- Fixed bug preventing calling MCP tools that have nested references in their input schemas
1124- Silenced a noisy but harmless error during upgrades
1125- Improved ultrathink text display
1126- Improved clarity of 5-hour session limit warning message
1127
1128## 2.0.49
1129
1130- Added readline-style ctrl-y for pasting deleted text
1131- Improved clarity of usage limit warning message
1132- Fixed handling of subagent permissions
1133
1134## 2.0.47
1135
1136- Improved error messages and validation for `claude --teleport`
1137- Improved error handling in `/usage`
1138- Fixed race condition with history entry not getting logged at exit
1139- Fixed Vertex AI configuration not being applied from `settings.json`
1140
1141## 2.0.46
1142
1143- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
1144
1145## 2.0.45
1146
1147- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
1148- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
1149- Send background tasks to Claude Code on the web by starting a message with `&`
1150
1151## 2.0.43
1152
1153- Added `permissionMode` field for custom agents
1154- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
1155- Added skills frontmatter field to declare skills to auto-load for subagents
1156- Added the `SubagentStart` hook event
1157- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
1158- Fixed duplicate rendering of some messages in the UI
1159- Fixed some visual flickers
1160- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
1161
1162## 2.0.42
1163
1164- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
1165
1166## 2.0.41
1167
1168- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
1169- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
1170- Fixed incorrect labeling of user settings vs project settings in command descriptions
1171- Fixed crash when plugin command hooks timeout during execution
1172- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
1173- Fixed broken security documentation links in trust dialogs and onboarding
1174- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
1175- ctrl-r history search landing on a slash command no longer cancels the search
1176- SDK: Support custom timeouts for hooks
1177- Allow more safe git commands to run without approval
1178- Plugins: Added support for sharing and installing output styles
1179- Teleporting a session from web will automatically set the upstream branch
1180
1181## 2.0.37
1182
1183- Fixed how idleness is computed for notifications
1184- Hooks: Added matcher values for Notification hook events
1185- Output Styles: Added `keep-coding-instructions` option to frontmatter
1186
1187## 2.0.36
1188
1189- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
1190- Fixed queued messages being incorrectly executed as bash commands
1191- Fixed input being lost when typing while a queued message is processed
1192
1193## 2.0.35
1194
1195- Improve fuzzy search results when searching commands
1196- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
1197- Added `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` environment variable to automatically exit SDK mode after a specified idle duration, useful for automated workflows and scripts
1198- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
1199- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
1200
1201## 2.0.34
1202
1203- VSCode Extension: Added setting to configure the initial permission mode for new conversations
1204- Improved file path suggestion performance with native Rust-based fuzzy finder
1205- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
1206- Fixed memory crash when reading or writing large files (especially base64-encoded images)
1207
1208## 2.0.33
1209
1210- Native binary installs now launch quicker.
1211- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
1212- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
1213
1214## 2.0.32
1215
1216- Un-deprecate output styles based on community feedback
1217- Added `companyAnnouncements` setting for displaying announcements on startup
1218- Fixed hook progress messages not updating correctly during PostToolUse hook execution
1219
1220## 2.0.31
1221
1222- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1223- Vertex: add support for Web Search on supported models
1224- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1225- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1226- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1227- Fixed plugin uninstall not removing plugins
1228
1229## 2.0.30
1230
1231- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1232- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1233- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1234- Added prompt-based stop hooks
1235- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1236- Enabled SSE MCP servers on native build
1237- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1238- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1239- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1240- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1241- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1242- Fixed bug that saved session permissions to local settings
1243- Fixed MCP tools not being available to sub-agents
1244- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1245- Fixed delay when navigating through typeahead suggestions with arrow keys
1246- VSCode: Restored selection indicator in input footer showing current file or code selection status
1247
1248## 2.0.28
1249
1250- Plan mode: introduced new Plan subagent
1251- Subagents: claude can now choose to resume subagents
1252- Subagents: claude can dynamically choose the model used by its subagents
1253- SDK: added --max-budget-usd flag
1254- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1255- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1256- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1257- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1258- Various other bug fixes
1259
1260## 2.0.27
1261
1262- New UI for permission prompts
1263- Added current branch filtering and search to session resume screen for easier navigation
1264- Fixed directory @-mention causing "No assistant message found" error
1265- VSCode Extension: Add config setting to include .gitignored files in file searches
1266- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1267
1268## 2.0.25
1269
1270- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://platform.claude.com/docs/en/agent-sdk/migration-guide
1271
1272## 2.0.24
1273
1274- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1275- Claude Code Web: Support for Web -> CLI teleport
1276- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1277- Bedrock: Display awsAuthRefresh output when auth is required
1278
1279## 2.0.22
1280
1281- Fixed content layout shift when scrolling through slash commands
1282- IDE: Add toggle to enable/disable thinking.
1283- Fix bug causing duplicate permission prompts with parallel tool calls
1284- Add support for enterprise managed MCP allowlist and denylist
1285
1286## 2.0.21
1287
1288- Support MCP `structuredContent` field in tool responses
1289- Added an interactive question tool
1290- Claude will now ask you questions more often in plan mode
1291- Added Haiku 4.5 as a model option for Pro users
1292- Fixed an issue where queued commands don't have access to previous messages' output
1293
1294## 2.0.20
1295
1296- Added support for Claude Skills
1297
1298## 2.0.19
1299
1300- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1301- Fixed a bug where Haiku was unnecessarily called in print mode
1302
1303## 2.0.17
1304
1305- Added Haiku 4.5 to model selector!
1306- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1307- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1308- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1309- OTEL: support HTTP_PROXY and HTTPS_PROXY
1310- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1311
1312## 2.0.15
1313
1314- Fixed bug with resuming where previously created files needed to be read again before writing
1315- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1316
1317## 2.0.14
1318
1319- Fix @-mentioning MCP servers to toggle them on/off
1320- Improve permission checks for bash with inline env vars
1321- Fix ultrathink + thinking toggle
1322- Reduce unnecessary logins
1323- Document --system-prompt
1324- Several improvements to rendering
1325- Plugins UI polish
1326
1327## 2.0.13
1328
1329- Fixed `/plugin` not working on native build
1330
1331## 2.0.12
1332
1333- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1334- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1335- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1336- `/plugin validate` command for validating plugin structure and configuration
1337- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1338- Plugin documentation available at https://code.claude.com/docs/en/plugins
1339- Comprehensive error messages and diagnostics via `/doctor` command
1340- Avoid flickering in `/model` selector
1341- Improvements to `/help`
1342- Avoid mentioning hooks in `/resume` summaries
1343- Changes to the "verbose" setting in `/config` now persist across sessions
1344
1345## 2.0.11
1346
1347- Reduced system prompt size by 1.4k tokens
1348- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1349- Fixed Opus fallback rate limit errors appearing incorrectly
1350- Fixed /add-dir command selecting wrong default tab
1351
1352## 2.0.10
1353
1354- Rewrote terminal renderer for buttery smooth UI
1355- Enable/disable MCP servers by @mentioning, or in /mcp
1356- Added tab completion for shell commands in bash mode
1357- PreToolUse hooks can now modify tool inputs
1358- Press Ctrl-G to edit your prompt in your system's configured text editor
1359- Fixes for bash permission checks with environment variables in the command
1360
1361## 2.0.9
1362
1363- Fix regression where bash backgrounding stopped working
1364
1365## 2.0.8
1366
1367- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1368- IDE: Add drag-and-drop support for files and folders in chat
1369- /context: Fix counting for thinking blocks
1370- Improve message rendering for users with light themes on dark terminals
1371- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1372
1373## 2.0.5
1374
1375- IDE: Fix IME unintended message submission with Enter and Tab
1376- IDE: Add "Open in Terminal" link in login screen
1377- Fix unhandled OAuth expiration 401 API errors
1378- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1379
1380## 2.0.1
1381
1382- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1383- Various bug fixes and presentation improvements
1384
1385## 2.0.0
1386
1387- New native VS Code extension
1388- Fresh coat of paint throughout the whole app
1389- /rewind a conversation to undo code changes
1390- /usage command to see plan limits
1391- Tab to toggle thinking (sticky across sessions)
1392- Ctrl-R to search history
1393- Unshipped claude config command
1394- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1395- SDK: The Claude Code SDK is now the Claude Agent SDK
1396- Add subagents dynamically with `--agents` flag
1397
1398## 1.0.126
1399
1400- Enable /context command for Bedrock and Vertex
1401- Add mTLS support for HTTP-based OpenTelemetry exporters
1402
1403## 1.0.124
1404
1405- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1406- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1407- No longer inform Claude of the list of allowed tools when permission is denied
1408- Fixed security vulnerability in Bash tool permission checks
1409- Improved VSCode extension performance for large files
1410
1411## 1.0.123
1412
1413- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1414- Fixed thinking mode triggering on negation phrases like "don't think"
1415- Fixed rendering performance degradation during token streaming
1416- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1417- Enhanced BashTool environment snapshot logging
1418- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1419- Migrated --debug logging to a file, to enable easy tailing & filtering
1420
1421## 1.0.120
1422
1423- Fix input lag during typing, especially noticeable with large prompts
1424- Improved VSCode extension command registry and sessions dialog user experience
1425- Enhanced sessions dialog responsiveness and visual feedback
1426- Fixed IDE compatibility issue by removing worktree support check
1427- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1428
1429## 1.0.119
1430
1431- Fix Windows issue where process visually freezes on entering interactive mode
1432- Support dynamic headers for MCP servers via headersHelper configuration
1433- Fix thinking mode not working in headless sessions
1434- Fix slash commands now properly update allowed tools instead of replacing them
1435
1436## 1.0.117
1437
1438- Add Ctrl-R history search to recall previous commands like bash/zsh
1439- Fix input lag while typing, especially on Windows
1440- Add sed command to auto-allowed commands in acceptEdits mode
1441- Fix Windows PATH comparison to be case-insensitive for drive letters
1442- Add permissions management hint to /add-dir output
1443
1444## 1.0.115
1445
1446- Improve thinking mode display with enhanced visual effects
1447- Type /t to temporarily disable thinking mode in your prompt
1448- Improve path validation for glob and grep tools
1449- Show condensed output for post-tool hooks to reduce visual clutter
1450- Fix visual feedback when loading state completes
1451- Improve UI consistency for permission request dialogs
1452
1453## 1.0.113
1454
1455- Deprecated piped input in interactive mode
1456- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1457
1458## 1.0.112
1459
1460- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1461- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1462- Hooks: Added systemMessage support for SessionEnd hooks
1463- Added `spinnerTipsEnabled` setting to disable spinner tips
1464- IDE: Various improvements and bug fixes
1465
1466## 1.0.111
1467
1468- /model now validates provided model names
1469- Fixed Bash tool crashes caused by malformed shell syntax parsing
1470
1471## 1.0.110
1472
1473- /terminal-setup command now supports WezTerm
1474- MCP: OAuth tokens now proactively refresh before expiration
1475- Fixed reliability issues with background Bash processes
1476
1477## 1.0.109
1478
1479- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1480
1481## 1.0.106
1482
1483- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1484
1485## 1.0.97
1486
1487- Settings: /doctor now validates permission rule syntax and suggests corrections
1488
1489## 1.0.94
1490
1491- Vertex: add support for global endpoints for supported models
1492- /memory command now allows direct editing of all imported memory files
1493- SDK: Add custom tools as callbacks
1494- Added /todos command to list current todo items
1495
1496## 1.0.93
1497
1498- Windows: Add alt + v shortcut for pasting images from clipboard
1499- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1500
1501## 1.0.90
1502
1503- Settings file changes take effect immediately - no restart required
1504
1505## 1.0.88
1506
1507- Fixed issue causing "OAuth authentication is currently not supported"
1508- Status line input now includes `exceeds_200k_tokens`
1509- Fixed incorrect usage tracking in /cost.
1510- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1511- Bedrock: Updated default Sonnet model to Sonnet 4
1512
1513## 1.0.86
1514
1515- Added /context to help users self-serve debug context issues
1516- SDK: Added UUID support for all SDK messages
1517- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1518
1519## 1.0.85
1520
1521- Status line input now includes session cost info
1522- Hooks: Introduced SessionEnd hook
1523
1524## 1.0.84
1525
1526- Fix tool_use/tool_result id mismatch error when network is unstable
1527- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1528- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1529- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1530
1531## 1.0.83
1532
1533- @-mention: Support files with spaces in path
1534- New shimmering spinner
1535
1536## 1.0.82
1537
1538- SDK: Add request cancellation support
1539- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1540- Settings: Validation prevents invalid fields in .claude/settings.json files
1541- MCP: Improve tool name consistency
1542- Bash: Fix crash when Claude tries to automatically read large files
1543
1544## 1.0.81
1545
1546- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1547- Agents: Fix custom agent loading when agent files are unparsable
1548
1549## 1.0.80
1550
1551- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1552
1553## 1.0.77
1554
1555- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1556- SDK: Add session support and permission denial tracking
1557- Fix token limit errors in conversation summarization
1558- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1559
1560## 1.0.73
1561
1562- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1563- MCP: Press Esc to cancel OAuth authentication flows
1564- Bash: Improved command validation and reduced false security warnings
1565- UI: Enhanced spinner animations and status line visual hierarchy
1566- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1567
1568## 1.0.72
1569
1570- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1571
1572## 1.0.71
1573
1574- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1575- Customizable status line: add your terminal prompt to Claude Code with /statusline
1576
1577## 1.0.70
1578
1579- Performance: Optimized message rendering for better performance with large contexts
1580- Windows: Fixed native file search, ripgrep, and subagent functionality
1581- Added support for @-mentions in slash command arguments
1582
1583## 1.0.69
1584
1585- Upgraded Opus to version 4.1
1586
1587## 1.0.68
1588
1589- Fix incorrect model names being used for certain commands like `/pr-comments`
1590- Windows: improve permissions checks for allow / deny tools and project trust. This may create a new project entry in `.claude.json` - manually merge the history field if desired.
1591- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1592- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1593- SDK: Added canUseTool callback support for tool confirmation
1594- Added `disableAllHooks` setting
1595- Improved file suggestions performance in large repos
1596
1597## 1.0.65
1598
1599- IDE: Fixed connection stability issues and error handling for diagnostics
1600- Windows: Fixed shell environment setup for users without .bashrc files
1601
1602## 1.0.64
1603
1604- Agents: Added model customization support - you can now specify which model an agent should use
1605- Agents: Fixed unintended access to the recursive agent tool
1606- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1607- SDK: Fixed user input tracking across multi-turn conversations
1608- Added hidden files to file search and @-mention suggestions
1609
1610## 1.0.63
1611
1612- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1613
1614## 1.0.62
1615
1616- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1617- Hooks: Added SessionStart hook for new session initialization
1618- /add-dir command now supports typeahead for directory paths
1619- Improved network connectivity check reliability
1620
1621## 1.0.61
1622
1623- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1624- Settings: Added `--settings` flag to load settings from a JSON file
1625- Settings: Fixed resolution of settings files paths that are symlinks
1626- OTEL: Fixed reporting of wrong organization after authentication changes
1627- Slash commands: Fixed permissions checking for allowed-tools with Bash
1628- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1629- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1630- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1631
1632## 1.0.60
1633
1634- You can now create custom subagents for specialized tasks! Run /agents to get started
1635
1636## 1.0.59
1637
1638- SDK: Added tool confirmation support with canUseTool callback
1639- SDK: Allow specifying env for spawned process
1640- Hooks: Exposed PermissionDecision to hooks (including "ask")
1641- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1642- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1643
1644## 1.0.58
1645
1646- Added support for reading PDFs
1647- MCP: Improved server health status display in 'claude mcp list'
1648- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1649
1650## 1.0.57
1651
1652- Added support for specifying a model in slash commands
1653- Improved permission messages to help Claude understand allowed tools
1654- Fix: Remove trailing newlines from bash output in terminal wrapping
1655
1656## 1.0.56
1657
1658- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1659- Fixes for WSL IDE detection
1660- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1661
1662## 1.0.55
1663
1664- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1665- Windows: fixed Ctrl+Z crash
1666- SDK: Added ability to capture error logging
1667- Add --system-prompt-file option to override system prompt in print mode
1668
1669## 1.0.54
1670
1671- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1672- Custom slash commands: Added argument-hint to frontmatter
1673- Windows: OAuth uses port 45454 and properly constructs browser URL
1674- Windows: mode switching now uses alt + m, and plan mode renders properly
1675- Shell: Switch to in-memory shell snapshot to fix file-related errors
1676
1677## 1.0.53
1678
1679- Updated @-mention file truncation from 100 lines to 2000 lines
1680- Add helper script settings for AWS token refresh: awsAuthRefresh (for foreground operations like aws sso login) and awsCredentialExport (for background operation with STS-like response).
1681
1682## 1.0.52
1683
1684- Added support for MCP server instructions
1685
1686## 1.0.51
1687
1688- Added support for native Windows (requires Git for Windows)
1689- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1690- Settings: /doctor can now help you identify and fix invalid setting files
1691- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1692- Increased auto-compact warning threshold from 60% to 80%
1693- Fixed an issue with handling user directories with spaces for shell snapshots
1694- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1695- Custom slash commands: Fixed user-level commands in subdirectories
1696- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1697
1698## 1.0.48
1699
1700- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1701- Added progress messages to Bash tool based on the last 5 lines of command output
1702- Added expanding variables support for MCP server configuration
1703- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1704- Improved IDE extension path handling when Claude Code runs in WSL
1705- Hooks: Added a PreCompact hook
1706- Vim mode: Added c, f/F, t/T
1707
1708## 1.0.45
1709
1710- Redesigned Search (Grep) tool with new tool input parameters and features
1711- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1712- Fixed config file corruption issue by enforcing atomic writes
1713- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1714- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1715- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1716
1717## 1.0.44
1718
1719- New /export command lets you quickly export a conversation for sharing
1720- MCP: resource_link tool results are now supported
1721- MCP: tool annotations and tool titles now display in /mcp view
1722- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1723
1724## 1.0.43
1725
1726- Fixed a bug where the theme selector was saving excessively
1727- Hooks: Added EPIPE system error handling
1728
1729## 1.0.42
1730
1731- Added tilde (`~`) expansion support to `/add-dir` command
1732
1733## 1.0.41
1734
1735- Hooks: Split Stop hook triggering into Stop and SubagentStop
1736- Hooks: Enabled optional timeout configuration for each command
1737- Hooks: Added "hook_event_name" to hook input
1738- Fixed a bug where MCP tools would display twice in tool list
1739- New tool parameters JSON for Bash tool in `tool_decision` event
1740
1741## 1.0.40
1742
1743- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1744
1745## 1.0.39
1746
1747- New Active Time metric in OpenTelemetry logging
1748
1749## 1.0.38
1750
1751- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1752
1753## 1.0.37
1754
1755- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1756
1757## 1.0.36
1758
1759- Web search now takes today's date into context
1760- Fixed a bug where stdio MCP servers were not terminating properly on exit
1761
1762## 1.0.35
1763
1764- Added support for MCP OAuth Authorization Server discovery
1765
1766## 1.0.34
1767
1768- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1769
1770## 1.0.33
1771
1772- Improved logging functionality with session ID support
1773- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1774- Improvements to plan mode
1775
1776## 1.0.32
1777
1778- Updated loopback config for litellm
1779- Added forceLoginMethod setting to bypass login selection screen
1780
1781## 1.0.31
1782
1783- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1784
1785## 1.0.30
1786
1787- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1788- Improved file path autocomplete with filename matching
1789- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1790- Enhanced jq regex support for complex filters with pipes and select
1791
1792## 1.0.29
1793
1794- Improved CJK character support in cursor navigation and rendering
1795
1796## 1.0.28
1797
1798- Slash commands: Fix selector display during history navigation
1799- Resizes images before upload to prevent API size limit errors
1800- Added XDG_CONFIG_HOME support to configuration directory
1801- Performance optimizations for memory usage
1802- New attributes (terminal.type, language) in OpenTelemetry logging
1803
1804## 1.0.27
1805
1806- Streamable HTTP MCP servers are now supported
1807- Remote MCP servers (SSE and HTTP) now support OAuth
1808- MCP resources can now be @-mentioned
1809- /resume slash command to switch conversations within Claude Code
1810
1811## 1.0.25
1812
1813- Slash commands: moved "project" and "user" prefixes to descriptions
1814- Slash commands: improved reliability for command discovery
1815- Improved support for Ghostty
1816- Improved web search reliability
1817
1818## 1.0.24
1819
1820- Improved /mcp output
1821- Fixed a bug where settings arrays got overwritten instead of merged
1822
1823## 1.0.23
1824
1825- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1826- Released Python SDK: pip install claude-code-sdk to get started
1827
1828## 1.0.22
1829
1830- SDK: Renamed `total_cost` to `total_cost_usd`
1831
1832## 1.0.21
1833
1834- Improved editing of files with tab-based indentation
1835- Fix for tool_use without matching tool_result errors
1836- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1837
1838## 1.0.18
1839
1840- Added --add-dir CLI argument for specifying additional working directories
1841- Added streaming input support without require -p flag
1842- Improved startup performance and session storage performance
1843- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1844- Added detailed MCP server tools display (/mcp)
1845- MCP authentication and permission improvements
1846- Added auto-reconnection for MCP SSE connections on disconnect
1847- Fixed issue where pasted content was lost when dialogs appeared
1848
1849## 1.0.17
1850
1851- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1852- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1853- MCP server list UI improvements
1854- Update Claude Code process title to display "claude" instead of "node"
1855
1856## 1.0.11
1857
1858- Claude Code can now also be used with a Claude Pro subscription
1859- Added /upgrade for smoother switching to Claude Max plans
1860- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1861- Improved shell configuration error handling
1862- Improved todo list handling during compaction
1863
1864## 1.0.10
1865
1866- Added markdown table support
1867- Improved streaming performance
1868
1869## 1.0.8
1870
1871- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1872- Increased default otel interval from 1s -> 5s
1873- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1874- Fixed a regression where search tools unnecessarily asked for permissions
1875- Added support for triggering thinking non-English languages
1876- Improved compacting UI
1877
1878## 1.0.7
1879
1880- Renamed /allowed-tools -> /permissions
1881- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1882- Deprecated claude config commands in favor of editing settings.json
1883- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1884- Improved error handling for /install-github-app
1885- Bugfixes, UI polish, and tool reliability improvements
1886
1887## 1.0.6
1888
1889- Improved edit reliability for tab-indented files
1890- Respect CLAUDE_CONFIG_DIR everywhere
1891- Reduced unnecessary tool permission prompts
1892- Added support for symlinks in @file typeahead
1893- Bugfixes, UI polish, and tool reliability improvements
1894
1895## 1.0.4
1896
1897- Fixed a bug where MCP tool errors weren't being parsed correctly
1898
1899## 1.0.1
1900
1901- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1902- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1903- Updated documentation links and OAuth process descriptions
1904
1905## 1.0.0
1906
1907- Claude Code is now generally available
1908- Introducing Sonnet 4 and Opus 4 models
1909
1910## 0.2.125
1911
1912- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1913- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1914
1915## 0.2.117
1916
1917- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1918- Introduced settings.cleanupPeriodDays
1919- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1920- Introduced --debug mode
1921
1922## 0.2.108
1923
1924- You can now send messages to Claude while it works to steer Claude in real-time
1925- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1926- Fixed a bug where thinking was not working in -p mode
1927- Fixed a regression in /cost reporting
1928- Deprecated MCP wizard interface in favor of other MCP commands
1929- Lots of other bugfixes and improvements
1930
1931## 0.2.107
1932
1933- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1934
1935## 0.2.106
1936
1937- MCP SSE server configs can now specify custom headers
1938- Fixed a bug where MCP permission prompt didn't always show correctly
1939
1940## 0.2.105
1941
1942- Claude can now search the web
1943- Moved system & account status to /status
1944- Added word movement keybindings for Vim
1945- Improved latency for startup, todo tool, and file edits
1946
1947## 0.2.102
1948
1949- Improved thinking triggering reliability
1950- Improved @mention reliability for images and folders
1951- You can now paste multiple large chunks into one prompt
1952
1953## 0.2.100
1954
1955- Fixed a crash caused by a stack overflow error
1956- Made db storage optional; missing db support disables --continue and --resume
1957
1958## 0.2.98
1959
1960- Fixed an issue where auto-compact was running twice
1961
1962## 0.2.96
1963
1964- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1965
1966## 0.2.93
1967
1968- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1969- Claude now has access to a Todo list that helps it stay on track and be more organized
1970
1971## 0.2.82
1972
1973- Added support for --disallowedTools
1974- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1975
1976## 0.2.75
1977
1978- Hit Enter to queue up additional messages while Claude is working
1979- Drag in or copy/paste image files directly into the prompt
1980- @-mention files to directly add them to context
1981- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1982- Improved performance for filename auto-complete
1983
1984## 0.2.74
1985
1986- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1987- Task tool can now perform writes and run bash commands
1988
1989## 0.2.72
1990
1991- Updated spinner to indicate tokens loaded and tool usage
1992
1993## 0.2.70
1994
1995- Network commands like curl are now available for Claude to use
1996- Claude can now run multiple web queries in parallel
1997- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1998
1999## 0.2.69
2000
2001- Fixed UI glitches with improved Select component behavior
2002- Enhanced terminal output display with better text truncation logic
2003
2004## 0.2.67
2005
2006- Shared project permission rules can be saved in .claude/settings.json
2007
2008## 0.2.66
2009
2010- Print mode (-p) now supports streaming output via --output-format=stream-json
2011- Fixed issue where pasting could trigger memory or bash mode unexpectedly
2012
2013## 0.2.63
2014
2015- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
2016
2017## 0.2.61
2018
2019- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
2020- Enhanced image detection for more reliable clipboard paste functionality
2021- Fixed an issue where ESC key could crash the conversation history selector
2022
2023## 0.2.59
2024
2025- Copy+paste images directly into your prompt
2026- Improved progress indicators for bash and fetch tools
2027- Bugfixes for non-interactive mode (-p)
2028
2029## 0.2.54
2030
2031- Quickly add to Memory by starting your message with '#'
2032- Press ctrl+r to see full output for long tool results
2033- Added support for MCP SSE transport
2034
2035## 0.2.53
2036
2037- New web fetch tool lets Claude view URLs that you paste in
2038- Fixed a bug with JPEG detection
2039
2040## 0.2.50
2041
2042- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
2043
2044## 0.2.49
2045
2046- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
2047
2048## 0.2.47
2049
2050- Press Tab to auto-complete file and folder names
2051- Press Shift + Tab to toggle auto-accept for file edits
2052- Automatic conversation compaction for infinite conversation length (toggle with /config)
2053
2054## 0.2.44
2055
2056- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
2057
2058## 0.2.41
2059
2060- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
2061- MCP server startup no longer blocks the app from starting up
2062
2063## 0.2.37
2064
2065- New /release-notes command lets you view release notes at any time
2066- `claude config add/remove` commands now accept multiple values separated by commas or spaces
2067
2068## 0.2.36
2069
2070- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
2071- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
2072
2073## 0.2.34
2074
2075- Vim bindings for text input - enable with /vim or /config
2076
2077## 0.2.32
2078
2079- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
2080- Fix for some PersistentShell issues
2081
2082## 0.2.31
2083
2084- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
2085- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
2086
2087## 0.2.30
2088
2089- Added ANSI color theme for better terminal compatibility
2090- Fixed issue where slash command arguments weren't being sent properly
2091- (Mac-only) API keys are now stored in macOS Keychain
2092
2093## 0.2.26
2094
2095- New /approved-tools command for managing tool permissions
2096- Word-level diff display for improved code readability
2097- Fuzzy matching for slash commands
2098
2099## 0.2.21
2100
2101- Fuzzy matching for /commands