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 a6a8045 · on Jan 20, 2026 History
1,443 lines · 71.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.14
4
5- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
6- Added search to installed plugins list - type to filter by name or description
7- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
8- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
9- Fixed memory issues that could cause crashes when running parallel subagents
10- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
11- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
12- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
13- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
14- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
15- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
16- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
17- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
18- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
19- Improved backspace to delete pasted text as a single token instead of one character at a time
20- [VSCode] Added `/usage` command to display current plan usage
21
22## 2.1.12
23
24- Fixed message rendering bug
25
26## 2.1.11
27
28- Fixed excessive MCP connection requests for HTTP/SSE transports
29
30## 2.1.10
31
32- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
33- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
34- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
35- Improved startup to capture keystrokes typed before the REPL is fully ready
36- Improved file suggestions to show as removable attachments instead of inserting text when accepted
37- [VSCode] Added install count display to plugin listings
38- [VSCode] Added trust warning when installing plugins
39
40## 2.1.9
41
42- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
43- Added `plansDirectory` setting to customize where plan files are stored
44- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
45- Added session URL attribution to commits and PRs created from web sessions
46- Added support for `PreToolUse` hooks to return `additionalContext` to the model
47- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
48- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
49- Fixed MCP server reconnection hanging when cached connection promise never resolves
50- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
51
52## 2.1.7
53
54- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
55- Added ability to provide feedback when accepting permission prompts
56- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
57- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
58- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
59- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
60- 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)
61- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
62- Fixed terminal title animation jitter by using fixed-width braille characters
63- Fixed plugins with git submodules not being fully initialized when installed
64- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
65- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
66- 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.
67- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
68- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
69
70## 2.1.6
71
72- Added search functionality to `/config` command for quickly filtering settings
73- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
74- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
75- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
76- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
77- Added an error display when the editor fails during Ctrl+G
78- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
79- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
80- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
81- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
82- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
83- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
84- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
85- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
86- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
87- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
88- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
89- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
90- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
91- Improved @ autocomplete with icons for different suggestion types and single-line formatting
92- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
93- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
94- Changed terminal title to "Claude Code" on startup for better window identification
95- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
96- [VSCode] Fixed usage indicator not updating after manual compact
97
98## 2.1.5
99
100- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
101
102## 2.1.4
103
104- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
105- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
106
107## 2.1.3
108
109- Merged slash commands and skills, simplifying the mental model with no change in behavior
110- Added release channel (`stable` or `latest`) toggle to `/config`
111- 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
112- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
113- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
114- Fixed mismatch between background task count in status bar and items shown in tasks dialog
115- Fixed sub-agents using the wrong model during conversation compaction
116- Fixed web search in sub-agents using incorrect model
117- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
118- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
119- Improved slash command suggestion readability by truncating long descriptions to 2 lines
120- Changed tool hook execution timeout from 60 seconds to 10 minutes
121- [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)
122
123## 2.1.2
124
125- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
126- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
127- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
128- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
129- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
130- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
131- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
132- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
133- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
134- Fixed updates incorrectly claiming another installation is in progress
135- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
136- Fixed remote session URL and teleport being broken when using `/tasks` command
137- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
138- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
139- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
140- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
141- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
142- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
143- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
144- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
145- [SDK] Changed minimum zod peer dependency to ^4.0.0
146- [VSCode] Fixed usage display not updating after manual compact
147
148## 2.1.0
149
150- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
151- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
152- Added support for `agent` field in skills to specify agent type for execution
153- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
154- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
155- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
156- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
157- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
158- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
159- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
160- Fixed Esc key with queued prompts to only move them to input without canceling the running task
161- Reduced permission prompts for complex bash commands
162- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
163- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
164- Fixed plugin path resolution for file-based marketplace sources
165- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
166- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
167- Fixed Claude in Chrome support for WSL environments
168- Fixed Windows native installer silently failing when executable creation fails
169- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
170- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
171- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
172- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
173- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
174- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
175- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
176- Added hooks support for skill and slash command frontmatter
177- 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
178- Added `/plan` command shortcut to enable plan mode directly from the prompt
179- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
180- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
181- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
182- Added support for `once: true` config for hooks
183- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
184- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
185- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
186- Added left/right arrow key navigation for cycling through tabs in dialogs
187- Added real-time thinking block display in Ctrl+O transcript mode
188- Added filepath to full output in background bash task details dialog
189- Added Skills as a separate category in the context visualization
190- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
191- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
192- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
193- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
194- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
195- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
196- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
197- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
198- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
199- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
200- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
201- Fixed commands with `$()` command substitution failing with parse errors
202- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
203- 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)
204- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
205- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
206- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
207- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
208- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
209- Fixed subagents sometimes not inheriting the parent's model by default
210- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
211- Fixed duplicate Bash commands appearing in permission request option labels
212- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
213- Fixed background task completion notifications to appear proactively with bullet point
214- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
215- Fixed cursor disappearing after dismissing permission dialogs
216- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
217- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
218- Fixed images being silently dropped when queueing messages while backgrounding a task
219- Fixed large pasted images failing with "Image was too large" error
220- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
221- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
222- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
223- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
224- Fixed spinner token counter to properly accumulate tokens from subagents during execution
225- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
226- Fixed race condition where LSP tool could return "no server available" during startup
227- Fixed feedback submission hanging indefinitely when network requests timeout
228- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
229- Fixed hook success message showing trailing colon when hook has no output
230- Multiple optimizations to improve startup performance
231- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
232- Improved performance when reading Jupyter notebooks with many cells
233- Improved reliability for piped input like `cat refactor.md | claude`
234- Improved reliability for AskQuestion tool
235- Improved sed in-place edit commands to render as file edits with diff preview
236- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
237- Improved compaction reliability
238- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
239- Improved skills to show progress while executing, displaying tool uses as they happen
240- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
241- Improved skill suggestions to prioritize recently and frequently used skills
242- Improved spinner feedback when waiting for the first response token
243- Improved token count display in spinner to include tokens from background agents
244- Improved incremental output for async agents to give the main thread more control and visibility
245- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
246- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
247- Improved macOS screenshot paste reliability with TIFF format support
248- Improved `/stats` output
249- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
250- Changed "Interrupted" message color from red to grey for a less alarming appearance
251- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
252- Removed underline styling from image reference links
253- [SDK] Changed minimum zod peer dependency to ^4.0.0
254- [VSCode] Added currently selected model name to the context menu
255- [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")
256- [VSCode] Fixed paragraph breaks not rendering in markdown content
257- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
258- [Windows] Fixed issue with improper rendering
259
260## 2.0.76
261
262- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
263
264## 2.0.75
265
266- Minor bugfixes
267
268## 2.0.74
269
270- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
271- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
272- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
273- Added syntax highlighting info to theme picker
274- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
275- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
276- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
277- Fixed a potential crash when syntax highlighting isn't initialized correctly
278- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
279- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
280- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
281- [Windows] Fixed issue with improper rendering
282- [VSCode] Added gift tag pictogram for year-end promotion message
283
284## 2.0.73
285
286- Added clickable `[Image #N]` links that open attached images in the default viewer
287- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
288- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
289- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
290- Fixed slow input history cycling and race condition that could overwrite text after message submission
291- Improved `/theme` command to open theme picker directly
292- Improved theme picker UI
293- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
294- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
295
296## 2.0.72
297
298- 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
299- Reduced terminal flickering
300- Added scannable QR code to mobile app tip for quick app downloads
301- Added loading indicator when resuming conversations for better feedback
302- Fixed `/context` command not respecting custom system prompts in non-interactive mode
303- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
304- Improved @ mention file suggestion speed (~3x faster in git repositories)
305- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
306- Improved settings validation errors to be more prominent
307- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
308
309## 2.0.71
310
311- Added /config toggle to enable/disable prompt suggestions
312- Added `/settings` as an alias for the `/config` command
313- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
314- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
315- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
316- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
317- New syntax highlighting engine for native build
318
319## 2.0.70
320
321- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
322- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
323- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
324- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
325- Fixed input being cleared when processing queued commands while the user was typing
326- Fixed prompt suggestions replacing typed input when pressing Tab
327- Fixed diff view not updating when terminal is resized
328- Improved memory usage by 3x for large conversations
329- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
330- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
331- Fix thinking mode toggle in /config not persisting correctly
332- Improve UI for file creation permission dialog
333
334## 2.0.69
335
336- Minor bugfixes
337
338## 2.0.68
339
340- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
341- Fixed a bug where disallowed MCP tools were visible to the model
342- Fixed an issue where steering messages could be lost while a subagent is working
343- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
344- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
345- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
346
347## 2.0.67
348
349- Thinking mode is now enabled by default for Opus 4.5
350- Thinking mode configuration has moved to /config
351- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
352- Show reason why autoupdater is disabled in `/doctor`
353- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
354- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
355- Fixed scroll position resetting after deleting a permission rule in `/permissions`
356- 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
357- Fixed `claude install --force` not bypassing stale lock files
358- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
359- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
360
361## 2.0.65
362
363- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
364- Added context window information to status line input
365- Added `fileSuggestion` setting for custom `@` file search commands
366- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
367- Fixed prompt not being saved to history when aborting a query with Escape
368- Fixed Read tool image handling to identify format from bytes instead of file extension
369
370## 2.0.64
371
372- Made auto-compacting instant
373- Agents and bash commands can run asynchronously and send messages to wake up the main agent
374- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
375- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
376- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
377- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
378- Fixed auto-loading .env when using native installer
379- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
380- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
381- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
382- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
383- Bedrock: Improve efficiency of token counting
384- Bedrock: Add support for `aws login` AWS Management Console credentials
385- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
386
387## 2.0.62
388
389- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
390- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
391- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
392- Fixed slash command selection not working when multiple commands share the same name
393- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
394- Fixed running versions getting removed because lock file incorrectly going stale
395- Fixed IDE diff tab not closing when rejecting file changes
396
397## 2.0.61
398
399- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
400
401## 2.0.60
402
403- Added background agent support. Agents run in the background while you work
404- Added --disable-slash-commands CLI flag to disable all slash commands
405- Added model name to "Co-Authored-By" commit messages
406- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
407- Updated Fetch to skip summarization for pre-approved websites
408- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
409
410## 2.0.59
411
412- Added --agent CLI flag to override the agent setting for the current session
413- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
414- VS Code: Fixed .claude.json config file being read from incorrect location
415
416## 2.0.58
417
418- Pro users now have access to Opus 4.5 as part of their subscription!
419- Fixed timer duration showing "11m 60s" instead of "12m 0s"
420- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
421
422## 2.0.57
423
424- Added feedback input when rejecting plans, allowing users to tell Claude what to change
425- VSCode: Added streaming message support for real-time response display
426
427## 2.0.56
428
429- Added setting to enable/disable terminal progress bar (OSC 9;4)
430- 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.
431
432## 2.0.55
433
434- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
435- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
436- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
437- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
438
439## 2.0.54
440
441- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
442- Fix issue with excessive iTerm notifications
443
444## 2.0.52
445
446- Fixed duplicate message display when starting Claude with a command line argument
447- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
448- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
449- Permit some uses of `$!` in bash commands
450
451## 2.0.51
452
453- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
454- Introducing Claude Code for Desktop: https://claude.com/download
455- 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
456- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
457- Plan Mode now builds more precise plans and executes more thoroughly
458- Usage limit notifications now easier to understand
459- Switched `/usage` back to "% used"
460- Fixed handling of thinking errors
461- Fixed performance regression
462
463## 2.0.50
464
465- Fixed bug preventing calling MCP tools that have nested references in their input schemas
466- Silenced a noisy but harmless error during upgrades
467- Improved ultrathink text display
468- Improved clarity of 5-hour session limit warning message
469
470## 2.0.49
471
472- Added readline-style ctrl-y for pasting deleted text
473- Improved clarity of usage limit warning message
474- Fixed handling of subagent permissions
475
476## 2.0.47
477
478- Improved error messages and validation for `claude --teleport`
479- Improved error handling in `/usage`
480- Fixed race condition with history entry not getting logged at exit
481- Fixed Vertex AI configuration not being applied from `settings.json`
482
483## 2.0.46
484
485- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
486
487## 2.0.45
488
489- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
490- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
491- Send background tasks to Claude Code on the web by starting a message with `&`
492
493## 2.0.43
494
495- Added `permissionMode` field for custom agents
496- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
497- Added skills frontmatter field to declare skills to auto-load for subagents
498- Added the `SubagentStart` hook event
499- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
500- Fixed duplicate rendering of some messages in the UI
501- Fixed some visual flickers
502- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
503
504## 2.0.42
505
506- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
507
508## 2.0.41
509
510- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
511- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
512- Fixed incorrect labeling of user settings vs project settings in command descriptions
513- Fixed crash when plugin command hooks timeout during execution
514- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
515- Fixed broken security documentation links in trust dialogs and onboarding
516- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
517- ctrl-r history search landing on a slash command no longer cancels the search
518- SDK: Support custom timeouts for hooks
519- Allow more safe git commands to run without approval
520- Plugins: Added support for sharing and installing output styles
521- Teleporting a session from web will automatically set the upstream branch
522
523## 2.0.37
524
525- Fixed how idleness is computed for notifications
526- Hooks: Added matcher values for Notification hook events
527- Output Styles: Added `keep-coding-instructions` option to frontmatter
528
529## 2.0.36
530
531- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
532- Fixed queued messages being incorrectly executed as bash commands
533- Fixed input being lost when typing while a queued message is processed
534
535## 2.0.35
536
537- Improve fuzzy search results when searching commands
538- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
539- 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
540- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
541- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
542
543## 2.0.34
544
545- VSCode Extension: Added setting to configure the initial permission mode for new conversations
546- Improved file path suggestion performance with native Rust-based fuzzy finder
547- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
548- Fixed memory crash when reading or writing large files (especially base64-encoded images)
549
550## 2.0.33
551
552- Native binary installs now launch quicker.
553- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
554- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
555
556## 2.0.32
557
558- Un-deprecate output styles based on community feedback
559- Added `companyAnnouncements` setting for displaying announcements on startup
560- Fixed hook progress messages not updating correctly during PostToolUse hook execution
561
562## 2.0.31
563
564- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
565- Vertex: add support for Web Search on supported models
566- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
567- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
568- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
569- Fixed plugin uninstall not removing plugins
570
571## 2.0.30
572
573- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
574- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
575- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
576- Added prompt-based stop hooks
577- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
578- Enabled SSE MCP servers on native build
579- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
580- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
581- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
582- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
583- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
584- Fixed bug that saved session permissions to local settings
585- Fixed MCP tools not being available to sub-agents
586- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
587- Fixed delay when navigating through typeahead suggestions with arrow keys
588- VSCode: Restored selection indicator in input footer showing current file or code selection status
589
590## 2.0.28
591
592- Plan mode: introduced new Plan subagent
593- Subagents: claude can now choose to resume subagents
594- Subagents: claude can dynamically choose the model used by its subagents
595- SDK: added --max-budget-usd flag
596- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
597- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
598- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
599- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
600- Various other bug fixes
601
602## 2.0.27
603
604- New UI for permission prompts
605- Added current branch filtering and search to session resume screen for easier navigation
606- Fixed directory @-mention causing "No assistant message found" error
607- VSCode Extension: Add config setting to include .gitignored files in file searches
608- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
609
610## 2.0.25
611
612- 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
613
614## 2.0.24
615
616- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
617- Claude Code Web: Support for Web -> CLI teleport
618- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
619- Bedrock: Display awsAuthRefresh output when auth is required
620
621## 2.0.22
622
623- Fixed content layout shift when scrolling through slash commands
624- IDE: Add toggle to enable/disable thinking.
625- Fix bug causing duplicate permission prompts with parallel tool calls
626- Add support for enterprise managed MCP allowlist and denylist
627
628## 2.0.21
629
630- Support MCP `structuredContent` field in tool responses
631- Added an interactive question tool
632- Claude will now ask you questions more often in plan mode
633- Added Haiku 4.5 as a model option for Pro users
634- Fixed an issue where queued commands don't have access to previous messages' output
635
636## 2.0.20
637
638- Added support for Claude Skills
639
640## 2.0.19
641
642- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
643- Fixed a bug where Haiku was unnecessarily called in print mode
644
645## 2.0.17
646
647- Added Haiku 4.5 to model selector!
648- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
649- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
650- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
651- OTEL: support HTTP_PROXY and HTTPS_PROXY
652- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
653
654## 2.0.15
655
656- Fixed bug with resuming where previously created files needed to be read again before writing
657- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
658
659## 2.0.14
660
661- Fix @-mentioning MCP servers to toggle them on/off
662- Improve permission checks for bash with inline env vars
663- Fix ultrathink + thinking toggle
664- Reduce unnecessary logins
665- Document --system-prompt
666- Several improvements to rendering
667- Plugins UI polish
668
669## 2.0.13
670
671- Fixed `/plugin` not working on native build
672
673## 2.0.12
674
675- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
676- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
677- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
678- `/plugin validate` command for validating plugin structure and configuration
679- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
680- Plugin documentation available at https://code.claude.com/docs/en/plugins
681- Comprehensive error messages and diagnostics via `/doctor` command
682- Avoid flickering in `/model` selector
683- Improvements to `/help`
684- Avoid mentioning hooks in `/resume` summaries
685- Changes to the "verbose" setting in `/config` now persist across sessions
686
687## 2.0.11
688
689- Reduced system prompt size by 1.4k tokens
690- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
691- Fixed Opus fallback rate limit errors appearing incorrectly
692- Fixed /add-dir command selecting wrong default tab
693
694## 2.0.10
695
696- Rewrote terminal renderer for buttery smooth UI
697- Enable/disable MCP servers by @mentioning, or in /mcp
698- Added tab completion for shell commands in bash mode
699- PreToolUse hooks can now modify tool inputs
700- Press Ctrl-G to edit your prompt in your system's configured text editor
701- Fixes for bash permission checks with environment variables in the command
702
703## 2.0.9
704
705- Fix regression where bash backgrounding stopped working
706
707## 2.0.8
708
709- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
710- IDE: Add drag-and-drop support for files and folders in chat
711- /context: Fix counting for thinking blocks
712- Improve message rendering for users with light themes on dark terminals
713- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
714
715## 2.0.5
716
717- IDE: Fix IME unintended message submission with Enter and Tab
718- IDE: Add "Open in Terminal" link in login screen
719- Fix unhandled OAuth expiration 401 API errors
720- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
721
722## 2.0.1
723
724- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
725- Various bug fixes and presentation improvements
726
727## 2.0.0
728
729- New native VS Code extension
730- Fresh coat of paint throughout the whole app
731- /rewind a conversation to undo code changes
732- /usage command to see plan limits
733- Tab to toggle thinking (sticky across sessions)
734- Ctrl-R to search history
735- Unshipped claude config command
736- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
737- SDK: The Claude Code SDK is now the Claude Agent SDK
738- Add subagents dynamically with `--agents` flag
739
740## 1.0.126
741
742- Enable /context command for Bedrock and Vertex
743- Add mTLS support for HTTP-based OpenTelemetry exporters
744
745## 1.0.124
746
747- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
748- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
749- No longer inform Claude of the list of allowed tools when permission is denied
750- Fixed security vulnerability in Bash tool permission checks
751- Improved VSCode extension performance for large files
752
753## 1.0.123
754
755- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
756- Fixed thinking mode triggering on negation phrases like "don't think"
757- Fixed rendering performance degradation during token streaming
758- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
759- Enhanced BashTool environment snapshot logging
760- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
761- Migrated --debug logging to a file, to enable easy tailing & filtering
762
763## 1.0.120
764
765- Fix input lag during typing, especially noticeable with large prompts
766- Improved VSCode extension command registry and sessions dialog user experience
767- Enhanced sessions dialog responsiveness and visual feedback
768- Fixed IDE compatibility issue by removing worktree support check
769- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
770
771## 1.0.119
772
773- Fix Windows issue where process visually freezes on entering interactive mode
774- Support dynamic headers for MCP servers via headersHelper configuration
775- Fix thinking mode not working in headless sessions
776- Fix slash commands now properly update allowed tools instead of replacing them
777
778## 1.0.117
779
780- Add Ctrl-R history search to recall previous commands like bash/zsh
781- Fix input lag while typing, especially on Windows
782- Add sed command to auto-allowed commands in acceptEdits mode
783- Fix Windows PATH comparison to be case-insensitive for drive letters
784- Add permissions management hint to /add-dir output
785
786## 1.0.115
787
788- Improve thinking mode display with enhanced visual effects
789- Type /t to temporarily disable thinking mode in your prompt
790- Improve path validation for glob and grep tools
791- Show condensed output for post-tool hooks to reduce visual clutter
792- Fix visual feedback when loading state completes
793- Improve UI consistency for permission request dialogs
794
795## 1.0.113
796
797- Deprecated piped input in interactive mode
798- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
799
800## 1.0.112
801
802- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
803- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
804- Hooks: Added systemMessage support for SessionEnd hooks
805- Added `spinnerTipsEnabled` setting to disable spinner tips
806- IDE: Various improvements and bug fixes
807
808## 1.0.111
809
810- /model now validates provided model names
811- Fixed Bash tool crashes caused by malformed shell syntax parsing
812
813## 1.0.110
814
815- /terminal-setup command now supports WezTerm
816- MCP: OAuth tokens now proactively refresh before expiration
817- Fixed reliability issues with background Bash processes
818
819## 1.0.109
820
821- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
822
823## 1.0.106
824
825- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
826
827## 1.0.97
828
829- Settings: /doctor now validates permission rule syntax and suggests corrections
830
831## 1.0.94
832
833- Vertex: add support for global endpoints for supported models
834- /memory command now allows direct editing of all imported memory files
835- SDK: Add custom tools as callbacks
836- Added /todos command to list current todo items
837
838## 1.0.93
839
840- Windows: Add alt + v shortcut for pasting images from clipboard
841- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
842
843## 1.0.90
844
845- Settings file changes take effect immediately - no restart required
846
847## 1.0.88
848
849- Fixed issue causing "OAuth authentication is currently not supported"
850- Status line input now includes `exceeds_200k_tokens`
851- Fixed incorrect usage tracking in /cost.
852- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
853- Bedrock: Updated default Sonnet model to Sonnet 4
854
855## 1.0.86
856
857- Added /context to help users self-serve debug context issues
858- SDK: Added UUID support for all SDK messages
859- SDK: Added `--replay-user-messages` to replay user messages back to stdout
860
861## 1.0.85
862
863- Status line input now includes session cost info
864- Hooks: Introduced SessionEnd hook
865
866## 1.0.84
867
868- Fix tool_use/tool_result id mismatch error when network is unstable
869- Fix Claude sometimes ignoring real-time steering when wrapping up a task
870- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
871- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
872
873## 1.0.83
874
875- @-mention: Support files with spaces in path
876- New shimmering spinner
877
878## 1.0.82
879
880- SDK: Add request cancellation support
881- SDK: New additionalDirectories option to search custom paths, improved slash command processing
882- Settings: Validation prevents invalid fields in .claude/settings.json files
883- MCP: Improve tool name consistency
884- Bash: Fix crash when Claude tries to automatically read large files
885
886## 1.0.81
887
888- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
889- Agents: Fix custom agent loading when agent files are unparsable
890
891## 1.0.80
892
893- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
894
895## 1.0.77
896
897- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
898- SDK: Add session support and permission denial tracking
899- Fix token limit errors in conversation summarization
900- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
901
902## 1.0.73
903
904- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
905- MCP: Press Esc to cancel OAuth authentication flows
906- Bash: Improved command validation and reduced false security warnings
907- UI: Enhanced spinner animations and status line visual hierarchy
908- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
909
910## 1.0.72
911
912- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
913
914## 1.0.71
915
916- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
917- Customizable status line: add your terminal prompt to Claude Code with /statusline
918
919## 1.0.70
920
921- Performance: Optimized message rendering for better performance with large contexts
922- Windows: Fixed native file search, ripgrep, and subagent functionality
923- Added support for @-mentions in slash command arguments
924
925## 1.0.69
926
927- Upgraded Opus to version 4.1
928
929## 1.0.68
930
931- Fix incorrect model names being used for certain commands like `/pr-comments`
932- 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.
933- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
934- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
935- SDK: Added canUseTool callback support for tool confirmation
936- Added `disableAllHooks` setting
937- Improved file suggestions performance in large repos
938
939## 1.0.65
940
941- IDE: Fixed connection stability issues and error handling for diagnostics
942- Windows: Fixed shell environment setup for users without .bashrc files
943
944## 1.0.64
945
946- Agents: Added model customization support - you can now specify which model an agent should use
947- Agents: Fixed unintended access to the recursive agent tool
948- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
949- SDK: Fixed user input tracking across multi-turn conversations
950- Added hidden files to file search and @-mention suggestions
951
952## 1.0.63
953
954- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
955
956## 1.0.62
957
958- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
959- Hooks: Added SessionStart hook for new session initialization
960- /add-dir command now supports typeahead for directory paths
961- Improved network connectivity check reliability
962
963## 1.0.61
964
965- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
966- Settings: Added `--settings` flag to load settings from a JSON file
967- Settings: Fixed resolution of settings files paths that are symlinks
968- OTEL: Fixed reporting of wrong organization after authentication changes
969- Slash commands: Fixed permissions checking for allowed-tools with Bash
970- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
971- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
972- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
973
974## 1.0.60
975
976- You can now create custom subagents for specialized tasks! Run /agents to get started
977
978## 1.0.59
979
980- SDK: Added tool confirmation support with canUseTool callback
981- SDK: Allow specifying env for spawned process
982- Hooks: Exposed PermissionDecision to hooks (including "ask")
983- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
984- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
985
986## 1.0.58
987
988- Added support for reading PDFs
989- MCP: Improved server health status display in 'claude mcp list'
990- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
991
992## 1.0.57
993
994- Added support for specifying a model in slash commands
995- Improved permission messages to help Claude understand allowed tools
996- Fix: Remove trailing newlines from bash output in terminal wrapping
997
998## 1.0.56
999
1000- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1001- Fixes for WSL IDE detection
1002- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1003
1004## 1.0.55
1005
1006- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1007- Windows: fixed Ctrl+Z crash
1008- SDK: Added ability to capture error logging
1009- Add --system-prompt-file option to override system prompt in print mode
1010
1011## 1.0.54
1012
1013- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1014- Custom slash commands: Added argument-hint to frontmatter
1015- Windows: OAuth uses port 45454 and properly constructs browser URL
1016- Windows: mode switching now uses alt + m, and plan mode renders properly
1017- Shell: Switch to in-memory shell snapshot to fix file-related errors
1018
1019## 1.0.53
1020
1021- Updated @-mention file truncation from 100 lines to 2000 lines
1022- 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).
1023
1024## 1.0.52
1025
1026- Added support for MCP server instructions
1027
1028## 1.0.51
1029
1030- Added support for native Windows (requires Git for Windows)
1031- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1032- Settings: /doctor can now help you identify and fix invalid setting files
1033- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1034- Increased auto-compact warning threshold from 60% to 80%
1035- Fixed an issue with handling user directories with spaces for shell snapshots
1036- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1037- Custom slash commands: Fixed user-level commands in subdirectories
1038- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1039
1040## 1.0.48
1041
1042- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1043- Added progress messages to Bash tool based on the last 5 lines of command output
1044- Added expanding variables support for MCP server configuration
1045- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1046- Improved IDE extension path handling when Claude Code runs in WSL
1047- Hooks: Added a PreCompact hook
1048- Vim mode: Added c, f/F, t/T
1049
1050## 1.0.45
1051
1052- Redesigned Search (Grep) tool with new tool input parameters and features
1053- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1054- Fixed config file corruption issue by enforcing atomic writes
1055- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1056- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1057- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1058
1059## 1.0.44
1060
1061- New /export command lets you quickly export a conversation for sharing
1062- MCP: resource_link tool results are now supported
1063- MCP: tool annotations and tool titles now display in /mcp view
1064- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1065
1066## 1.0.43
1067
1068- Fixed a bug where the theme selector was saving excessively
1069- Hooks: Added EPIPE system error handling
1070
1071## 1.0.42
1072
1073- Added tilde (`~`) expansion support to `/add-dir` command
1074
1075## 1.0.41
1076
1077- Hooks: Split Stop hook triggering into Stop and SubagentStop
1078- Hooks: Enabled optional timeout configuration for each command
1079- Hooks: Added "hook_event_name" to hook input
1080- Fixed a bug where MCP tools would display twice in tool list
1081- New tool parameters JSON for Bash tool in `tool_decision` event
1082
1083## 1.0.40
1084
1085- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1086
1087## 1.0.39
1088
1089- New Active Time metric in OpenTelemetry logging
1090
1091## 1.0.38
1092
1093- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1094
1095## 1.0.37
1096
1097- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1098
1099## 1.0.36
1100
1101- Web search now takes today's date into context
1102- Fixed a bug where stdio MCP servers were not terminating properly on exit
1103
1104## 1.0.35
1105
1106- Added support for MCP OAuth Authorization Server discovery
1107
1108## 1.0.34
1109
1110- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1111
1112## 1.0.33
1113
1114- Improved logging functionality with session ID support
1115- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1116- Improvements to plan mode
1117
1118## 1.0.32
1119
1120- Updated loopback config for litellm
1121- Added forceLoginMethod setting to bypass login selection screen
1122
1123## 1.0.31
1124
1125- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1126
1127## 1.0.30
1128
1129- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1130- Improved file path autocomplete with filename matching
1131- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1132- Enhanced jq regex support for complex filters with pipes and select
1133
1134## 1.0.29
1135
1136- Improved CJK character support in cursor navigation and rendering
1137
1138## 1.0.28
1139
1140- Slash commands: Fix selector display during history navigation
1141- Resizes images before upload to prevent API size limit errors
1142- Added XDG_CONFIG_HOME support to configuration directory
1143- Performance optimizations for memory usage
1144- New attributes (terminal.type, language) in OpenTelemetry logging
1145
1146## 1.0.27
1147
1148- Streamable HTTP MCP servers are now supported
1149- Remote MCP servers (SSE and HTTP) now support OAuth
1150- MCP resources can now be @-mentioned
1151- /resume slash command to switch conversations within Claude Code
1152
1153## 1.0.25
1154
1155- Slash commands: moved "project" and "user" prefixes to descriptions
1156- Slash commands: improved reliability for command discovery
1157- Improved support for Ghostty
1158- Improved web search reliability
1159
1160## 1.0.24
1161
1162- Improved /mcp output
1163- Fixed a bug where settings arrays got overwritten instead of merged
1164
1165## 1.0.23
1166
1167- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1168- Released Python SDK: pip install claude-code-sdk to get started
1169
1170## 1.0.22
1171
1172- SDK: Renamed `total_cost` to `total_cost_usd`
1173
1174## 1.0.21
1175
1176- Improved editing of files with tab-based indentation
1177- Fix for tool_use without matching tool_result errors
1178- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1179
1180## 1.0.18
1181
1182- Added --add-dir CLI argument for specifying additional working directories
1183- Added streaming input support without require -p flag
1184- Improved startup performance and session storage performance
1185- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1186- Added detailed MCP server tools display (/mcp)
1187- MCP authentication and permission improvements
1188- Added auto-reconnection for MCP SSE connections on disconnect
1189- Fixed issue where pasted content was lost when dialogs appeared
1190
1191## 1.0.17
1192
1193- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1194- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1195- MCP server list UI improvements
1196- Update Claude Code process title to display "claude" instead of "node"
1197
1198## 1.0.11
1199
1200- Claude Code can now also be used with a Claude Pro subscription
1201- Added /upgrade for smoother switching to Claude Max plans
1202- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1203- Improved shell configuration error handling
1204- Improved todo list handling during compaction
1205
1206## 1.0.10
1207
1208- Added markdown table support
1209- Improved streaming performance
1210
1211## 1.0.8
1212
1213- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1214- Increased default otel interval from 1s -> 5s
1215- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1216- Fixed a regression where search tools unnecessarily asked for permissions
1217- Added support for triggering thinking non-English languages
1218- Improved compacting UI
1219
1220## 1.0.7
1221
1222- Renamed /allowed-tools -> /permissions
1223- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1224- Deprecated claude config commands in favor of editing settings.json
1225- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1226- Improved error handling for /install-github-app
1227- Bugfixes, UI polish, and tool reliability improvements
1228
1229## 1.0.6
1230
1231- Improved edit reliability for tab-indented files
1232- Respect CLAUDE_CONFIG_DIR everywhere
1233- Reduced unnecessary tool permission prompts
1234- Added support for symlinks in @file typeahead
1235- Bugfixes, UI polish, and tool reliability improvements
1236
1237## 1.0.4
1238
1239- Fixed a bug where MCP tool errors weren't being parsed correctly
1240
1241## 1.0.1
1242
1243- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1244- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1245- Updated documentation links and OAuth process descriptions
1246
1247## 1.0.0
1248
1249- Claude Code is now generally available
1250- Introducing Sonnet 4 and Opus 4 models
1251
1252## 0.2.125
1253
1254- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1255- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1256
1257## 0.2.117
1258
1259- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1260- Introduced settings.cleanupPeriodDays
1261- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1262- Introduced --debug mode
1263
1264## 0.2.108
1265
1266- You can now send messages to Claude while it works to steer Claude in real-time
1267- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1268- Fixed a bug where thinking was not working in -p mode
1269- Fixed a regression in /cost reporting
1270- Deprecated MCP wizard interface in favor of other MCP commands
1271- Lots of other bugfixes and improvements
1272
1273## 0.2.107
1274
1275- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1276
1277## 0.2.106
1278
1279- MCP SSE server configs can now specify custom headers
1280- Fixed a bug where MCP permission prompt didn't always show correctly
1281
1282## 0.2.105
1283
1284- Claude can now search the web
1285- Moved system & account status to /status
1286- Added word movement keybindings for Vim
1287- Improved latency for startup, todo tool, and file edits
1288
1289## 0.2.102
1290
1291- Improved thinking triggering reliability
1292- Improved @mention reliability for images and folders
1293- You can now paste multiple large chunks into one prompt
1294
1295## 0.2.100
1296
1297- Fixed a crash caused by a stack overflow error
1298- Made db storage optional; missing db support disables --continue and --resume
1299
1300## 0.2.98
1301
1302- Fixed an issue where auto-compact was running twice
1303
1304## 0.2.96
1305
1306- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1307
1308## 0.2.93
1309
1310- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1311- Claude now has access to a Todo list that helps it stay on track and be more organized
1312
1313## 0.2.82
1314
1315- Added support for --disallowedTools
1316- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1317
1318## 0.2.75
1319
1320- Hit Enter to queue up additional messages while Claude is working
1321- Drag in or copy/paste image files directly into the prompt
1322- @-mention files to directly add them to context
1323- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1324- Improved performance for filename auto-complete
1325
1326## 0.2.74
1327
1328- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1329- Task tool can now perform writes and run bash commands
1330
1331## 0.2.72
1332
1333- Updated spinner to indicate tokens loaded and tool usage
1334
1335## 0.2.70
1336
1337- Network commands like curl are now available for Claude to use
1338- Claude can now run multiple web queries in parallel
1339- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1340
1341## 0.2.69
1342
1343- Fixed UI glitches with improved Select component behavior
1344- Enhanced terminal output display with better text truncation logic
1345
1346## 0.2.67
1347
1348- Shared project permission rules can be saved in .claude/settings.json
1349
1350## 0.2.66
1351
1352- Print mode (-p) now supports streaming output via --output-format=stream-json
1353- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1354
1355## 0.2.63
1356
1357- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1358
1359## 0.2.61
1360
1361- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1362- Enhanced image detection for more reliable clipboard paste functionality
1363- Fixed an issue where ESC key could crash the conversation history selector
1364
1365## 0.2.59
1366
1367- Copy+paste images directly into your prompt
1368- Improved progress indicators for bash and fetch tools
1369- Bugfixes for non-interactive mode (-p)
1370
1371## 0.2.54
1372
1373- Quickly add to Memory by starting your message with '#'
1374- Press ctrl+r to see full output for long tool results
1375- Added support for MCP SSE transport
1376
1377## 0.2.53
1378
1379- New web fetch tool lets Claude view URLs that you paste in
1380- Fixed a bug with JPEG detection
1381
1382## 0.2.50
1383
1384- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1385
1386## 0.2.49
1387
1388- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1389
1390## 0.2.47
1391
1392- Press Tab to auto-complete file and folder names
1393- Press Shift + Tab to toggle auto-accept for file edits
1394- Automatic conversation compaction for infinite conversation length (toggle with /config)
1395
1396## 0.2.44
1397
1398- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1399
1400## 0.2.41
1401
1402- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1403- MCP server startup no longer blocks the app from starting up
1404
1405## 0.2.37
1406
1407- New /release-notes command lets you view release notes at any time
1408- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1409
1410## 0.2.36
1411
1412- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1413- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1414
1415## 0.2.34
1416
1417- Vim bindings for text input - enable with /vim or /config
1418
1419## 0.2.32
1420
1421- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1422- Fix for some PersistentShell issues
1423
1424## 0.2.31
1425
1426- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1427- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1428
1429## 0.2.30
1430
1431- Added ANSI color theme for better terminal compatibility
1432- Fixed issue where slash command arguments weren't being sent properly
1433- (Mac-only) API keys are now stored in macOS Keychain
1434
1435## 0.2.26
1436
1437- New /approved-tools command for managing tool permissions
1438- Word-level diff display for improved code readability
1439- Fuzzy matching for slash commands
1440
1441## 0.2.21
1442
1443- Fuzzy matching for /commands