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 e9a9efc · on Jan 23, 2026 History
1,485 lines · 74.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.19
4
5- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
6- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
7- Fixed crashes on processors without AVX instruction support
8- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
9- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
10- Fixed resuming sessions by custom title not working when run from a different directory
11- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
12- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
13- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
14- Fixed file write preview omitting empty lines
15- Changed skills without additional permissions or hooks to be allowed without requiring approval
16- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
17- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
18- [VSCode] Enabled session forking and rewind functionality for all users
19
20## 2.1.18
21
22- 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
23
24## 2.1.17
25
26- Fixed crashes on processors without AVX instruction support
27
28## 2.1.16
29
30- Added new task management system, including new capabilities like dependency tracking
31- [VSCode] Added native plugin management support
32- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
33- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
34- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
35- Fixed session titles on the resume screen not respecting the user's language setting
36- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
37
38## 2.1.15
39
40- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
41- Improved UI rendering performance with React Compiler
42- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
43- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
44
45## 2.1.14
46
47- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
48- Added search to installed plugins list - type to filter by name or description
49- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
50- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
51- Fixed memory issues that could cause crashes when running parallel subagents
52- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
53- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
54- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
55- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
56- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
57- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
58- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
59- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
60- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
61- Improved backspace to delete pasted text as a single token instead of one character at a time
62- [VSCode] Added `/usage` command to display current plan usage
63
64## 2.1.12
65
66- Fixed message rendering bug
67
68## 2.1.11
69
70- Fixed excessive MCP connection requests for HTTP/SSE transports
71
72## 2.1.10
73
74- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
75- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
76- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
77- Improved startup to capture keystrokes typed before the REPL is fully ready
78- Improved file suggestions to show as removable attachments instead of inserting text when accepted
79- [VSCode] Added install count display to plugin listings
80- [VSCode] Added trust warning when installing plugins
81
82## 2.1.9
83
84- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
85- Added `plansDirectory` setting to customize where plan files are stored
86- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
87- Added session URL attribution to commits and PRs created from web sessions
88- Added support for `PreToolUse` hooks to return `additionalContext` to the model
89- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
90- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
91- Fixed MCP server reconnection hanging when cached connection promise never resolves
92- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
93
94## 2.1.7
95
96- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
97- Added ability to provide feedback when accepting permission prompts
98- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
99- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
100- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
101- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
102- 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)
103- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
104- Fixed terminal title animation jitter by using fixed-width braille characters
105- Fixed plugins with git submodules not being fully initialized when installed
106- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
107- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
108- 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.
109- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
110- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
111
112## 2.1.6
113
114- Added search functionality to `/config` command for quickly filtering settings
115- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
116- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
117- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
118- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
119- Added an error display when the editor fails during Ctrl+G
120- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
121- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
122- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
123- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
124- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
125- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
126- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
127- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
128- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
129- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
130- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
131- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
132- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
133- Improved @ autocomplete with icons for different suggestion types and single-line formatting
134- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
135- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
136- Changed terminal title to "Claude Code" on startup for better window identification
137- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
138- [VSCode] Fixed usage indicator not updating after manual compact
139
140## 2.1.5
141
142- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
143
144## 2.1.4
145
146- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
147- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
148
149## 2.1.3
150
151- Merged slash commands and skills, simplifying the mental model with no change in behavior
152- Added release channel (`stable` or `latest`) toggle to `/config`
153- 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
154- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
155- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
156- Fixed mismatch between background task count in status bar and items shown in tasks dialog
157- Fixed sub-agents using the wrong model during conversation compaction
158- Fixed web search in sub-agents using incorrect model
159- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
160- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
161- Improved slash command suggestion readability by truncating long descriptions to 2 lines
162- Changed tool hook execution timeout from 60 seconds to 10 minutes
163- [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)
164
165## 2.1.2
166
167- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
168- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
169- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
170- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
171- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
172- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
173- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
174- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
175- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
176- Fixed updates incorrectly claiming another installation is in progress
177- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
178- Fixed remote session URL and teleport being broken when using `/tasks` command
179- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
180- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
181- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
182- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
183- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
184- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
185- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
186- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
187- [SDK] Changed minimum zod peer dependency to ^4.0.0
188- [VSCode] Fixed usage display not updating after manual compact
189
190## 2.1.0
191
192- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
193- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
194- Added support for `agent` field in skills to specify agent type for execution
195- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
196- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
197- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
198- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
199- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
200- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
201- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
202- Fixed Esc key with queued prompts to only move them to input without canceling the running task
203- Reduced permission prompts for complex bash commands
204- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
205- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
206- Fixed plugin path resolution for file-based marketplace sources
207- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
208- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
209- Fixed Claude in Chrome support for WSL environments
210- Fixed Windows native installer silently failing when executable creation fails
211- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
212- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
213- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
214- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
215- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
216- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
217- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
218- Added hooks support for skill and slash command frontmatter
219- 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
220- Added `/plan` command shortcut to enable plan mode directly from the prompt
221- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
222- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
223- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
224- Added support for `once: true` config for hooks
225- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
226- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
227- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
228- Added left/right arrow key navigation for cycling through tabs in dialogs
229- Added real-time thinking block display in Ctrl+O transcript mode
230- Added filepath to full output in background bash task details dialog
231- Added Skills as a separate category in the context visualization
232- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
233- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
234- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
235- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
236- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
237- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
238- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
239- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
240- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
241- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
242- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
243- Fixed commands with `$()` command substitution failing with parse errors
244- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
245- 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)
246- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
247- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
248- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
249- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
250- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
251- Fixed subagents sometimes not inheriting the parent's model by default
252- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
253- Fixed duplicate Bash commands appearing in permission request option labels
254- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
255- Fixed background task completion notifications to appear proactively with bullet point
256- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
257- Fixed cursor disappearing after dismissing permission dialogs
258- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
259- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
260- Fixed images being silently dropped when queueing messages while backgrounding a task
261- Fixed large pasted images failing with "Image was too large" error
262- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
263- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
264- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
265- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
266- Fixed spinner token counter to properly accumulate tokens from subagents during execution
267- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
268- Fixed race condition where LSP tool could return "no server available" during startup
269- Fixed feedback submission hanging indefinitely when network requests timeout
270- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
271- Fixed hook success message showing trailing colon when hook has no output
272- Multiple optimizations to improve startup performance
273- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
274- Improved performance when reading Jupyter notebooks with many cells
275- Improved reliability for piped input like `cat refactor.md | claude`
276- Improved reliability for AskQuestion tool
277- Improved sed in-place edit commands to render as file edits with diff preview
278- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
279- Improved compaction reliability
280- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
281- Improved skills to show progress while executing, displaying tool uses as they happen
282- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
283- Improved skill suggestions to prioritize recently and frequently used skills
284- Improved spinner feedback when waiting for the first response token
285- Improved token count display in spinner to include tokens from background agents
286- Improved incremental output for async agents to give the main thread more control and visibility
287- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
288- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
289- Improved macOS screenshot paste reliability with TIFF format support
290- Improved `/stats` output
291- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
292- Changed "Interrupted" message color from red to grey for a less alarming appearance
293- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
294- Removed underline styling from image reference links
295- [SDK] Changed minimum zod peer dependency to ^4.0.0
296- [VSCode] Added currently selected model name to the context menu
297- [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")
298- [VSCode] Fixed paragraph breaks not rendering in markdown content
299- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
300- [Windows] Fixed issue with improper rendering
301
302## 2.0.76
303
304- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
305
306## 2.0.75
307
308- Minor bugfixes
309
310## 2.0.74
311
312- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
313- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
314- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
315- Added syntax highlighting info to theme picker
316- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
317- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
318- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
319- Fixed a potential crash when syntax highlighting isn't initialized correctly
320- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
321- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
322- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
323- [Windows] Fixed issue with improper rendering
324- [VSCode] Added gift tag pictogram for year-end promotion message
325
326## 2.0.73
327
328- Added clickable `[Image #N]` links that open attached images in the default viewer
329- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
330- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
331- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
332- Fixed slow input history cycling and race condition that could overwrite text after message submission
333- Improved `/theme` command to open theme picker directly
334- Improved theme picker UI
335- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
336- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
337
338## 2.0.72
339
340- 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
341- Reduced terminal flickering
342- Added scannable QR code to mobile app tip for quick app downloads
343- Added loading indicator when resuming conversations for better feedback
344- Fixed `/context` command not respecting custom system prompts in non-interactive mode
345- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
346- Improved @ mention file suggestion speed (~3x faster in git repositories)
347- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
348- Improved settings validation errors to be more prominent
349- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
350
351## 2.0.71
352
353- Added /config toggle to enable/disable prompt suggestions
354- Added `/settings` as an alias for the `/config` command
355- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
356- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
357- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
358- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
359- New syntax highlighting engine for native build
360
361## 2.0.70
362
363- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
364- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
365- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
366- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
367- Fixed input being cleared when processing queued commands while the user was typing
368- Fixed prompt suggestions replacing typed input when pressing Tab
369- Fixed diff view not updating when terminal is resized
370- Improved memory usage by 3x for large conversations
371- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
372- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
373- Fix thinking mode toggle in /config not persisting correctly
374- Improve UI for file creation permission dialog
375
376## 2.0.69
377
378- Minor bugfixes
379
380## 2.0.68
381
382- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
383- Fixed a bug where disallowed MCP tools were visible to the model
384- Fixed an issue where steering messages could be lost while a subagent is working
385- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
386- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
387- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
388
389## 2.0.67
390
391- Thinking mode is now enabled by default for Opus 4.5
392- Thinking mode configuration has moved to /config
393- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
394- Show reason why autoupdater is disabled in `/doctor`
395- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
396- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
397- Fixed scroll position resetting after deleting a permission rule in `/permissions`
398- 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
399- Fixed `claude install --force` not bypassing stale lock files
400- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
401- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
402
403## 2.0.65
404
405- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
406- Added context window information to status line input
407- Added `fileSuggestion` setting for custom `@` file search commands
408- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
409- Fixed prompt not being saved to history when aborting a query with Escape
410- Fixed Read tool image handling to identify format from bytes instead of file extension
411
412## 2.0.64
413
414- Made auto-compacting instant
415- Agents and bash commands can run asynchronously and send messages to wake up the main agent
416- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
417- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
418- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
419- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
420- Fixed auto-loading .env when using native installer
421- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
422- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
423- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
424- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
425- Bedrock: Improve efficiency of token counting
426- Bedrock: Add support for `aws login` AWS Management Console credentials
427- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
428
429## 2.0.62
430
431- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
432- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
433- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
434- Fixed slash command selection not working when multiple commands share the same name
435- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
436- Fixed running versions getting removed because lock file incorrectly going stale
437- Fixed IDE diff tab not closing when rejecting file changes
438
439## 2.0.61
440
441- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
442
443## 2.0.60
444
445- Added background agent support. Agents run in the background while you work
446- Added --disable-slash-commands CLI flag to disable all slash commands
447- Added model name to "Co-Authored-By" commit messages
448- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
449- Updated Fetch to skip summarization for pre-approved websites
450- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
451
452## 2.0.59
453
454- Added --agent CLI flag to override the agent setting for the current session
455- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
456- VS Code: Fixed .claude.json config file being read from incorrect location
457
458## 2.0.58
459
460- Pro users now have access to Opus 4.5 as part of their subscription!
461- Fixed timer duration showing "11m 60s" instead of "12m 0s"
462- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
463
464## 2.0.57
465
466- Added feedback input when rejecting plans, allowing users to tell Claude what to change
467- VSCode: Added streaming message support for real-time response display
468
469## 2.0.56
470
471- Added setting to enable/disable terminal progress bar (OSC 9;4)
472- 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.
473
474## 2.0.55
475
476- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
477- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
478- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
479- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
480
481## 2.0.54
482
483- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
484- Fix issue with excessive iTerm notifications
485
486## 2.0.52
487
488- Fixed duplicate message display when starting Claude with a command line argument
489- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
490- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
491- Permit some uses of `$!` in bash commands
492
493## 2.0.51
494
495- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
496- Introducing Claude Code for Desktop: https://claude.com/download
497- 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
498- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
499- Plan Mode now builds more precise plans and executes more thoroughly
500- Usage limit notifications now easier to understand
501- Switched `/usage` back to "% used"
502- Fixed handling of thinking errors
503- Fixed performance regression
504
505## 2.0.50
506
507- Fixed bug preventing calling MCP tools that have nested references in their input schemas
508- Silenced a noisy but harmless error during upgrades
509- Improved ultrathink text display
510- Improved clarity of 5-hour session limit warning message
511
512## 2.0.49
513
514- Added readline-style ctrl-y for pasting deleted text
515- Improved clarity of usage limit warning message
516- Fixed handling of subagent permissions
517
518## 2.0.47
519
520- Improved error messages and validation for `claude --teleport`
521- Improved error handling in `/usage`
522- Fixed race condition with history entry not getting logged at exit
523- Fixed Vertex AI configuration not being applied from `settings.json`
524
525## 2.0.46
526
527- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
528
529## 2.0.45
530
531- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
532- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
533- Send background tasks to Claude Code on the web by starting a message with `&`
534
535## 2.0.43
536
537- Added `permissionMode` field for custom agents
538- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
539- Added skills frontmatter field to declare skills to auto-load for subagents
540- Added the `SubagentStart` hook event
541- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
542- Fixed duplicate rendering of some messages in the UI
543- Fixed some visual flickers
544- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
545
546## 2.0.42
547
548- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
549
550## 2.0.41
551
552- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
553- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
554- Fixed incorrect labeling of user settings vs project settings in command descriptions
555- Fixed crash when plugin command hooks timeout during execution
556- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
557- Fixed broken security documentation links in trust dialogs and onboarding
558- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
559- ctrl-r history search landing on a slash command no longer cancels the search
560- SDK: Support custom timeouts for hooks
561- Allow more safe git commands to run without approval
562- Plugins: Added support for sharing and installing output styles
563- Teleporting a session from web will automatically set the upstream branch
564
565## 2.0.37
566
567- Fixed how idleness is computed for notifications
568- Hooks: Added matcher values for Notification hook events
569- Output Styles: Added `keep-coding-instructions` option to frontmatter
570
571## 2.0.36
572
573- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
574- Fixed queued messages being incorrectly executed as bash commands
575- Fixed input being lost when typing while a queued message is processed
576
577## 2.0.35
578
579- Improve fuzzy search results when searching commands
580- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
581- 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
582- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
583- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
584
585## 2.0.34
586
587- VSCode Extension: Added setting to configure the initial permission mode for new conversations
588- Improved file path suggestion performance with native Rust-based fuzzy finder
589- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
590- Fixed memory crash when reading or writing large files (especially base64-encoded images)
591
592## 2.0.33
593
594- Native binary installs now launch quicker.
595- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
596- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
597
598## 2.0.32
599
600- Un-deprecate output styles based on community feedback
601- Added `companyAnnouncements` setting for displaying announcements on startup
602- Fixed hook progress messages not updating correctly during PostToolUse hook execution
603
604## 2.0.31
605
606- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
607- Vertex: add support for Web Search on supported models
608- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
609- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
610- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
611- Fixed plugin uninstall not removing plugins
612
613## 2.0.30
614
615- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
616- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
617- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
618- Added prompt-based stop hooks
619- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
620- Enabled SSE MCP servers on native build
621- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
622- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
623- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
624- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
625- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
626- Fixed bug that saved session permissions to local settings
627- Fixed MCP tools not being available to sub-agents
628- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
629- Fixed delay when navigating through typeahead suggestions with arrow keys
630- VSCode: Restored selection indicator in input footer showing current file or code selection status
631
632## 2.0.28
633
634- Plan mode: introduced new Plan subagent
635- Subagents: claude can now choose to resume subagents
636- Subagents: claude can dynamically choose the model used by its subagents
637- SDK: added --max-budget-usd flag
638- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
639- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
640- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
641- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
642- Various other bug fixes
643
644## 2.0.27
645
646- New UI for permission prompts
647- Added current branch filtering and search to session resume screen for easier navigation
648- Fixed directory @-mention causing "No assistant message found" error
649- VSCode Extension: Add config setting to include .gitignored files in file searches
650- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
651
652## 2.0.25
653
654- 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
655
656## 2.0.24
657
658- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
659- Claude Code Web: Support for Web -> CLI teleport
660- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
661- Bedrock: Display awsAuthRefresh output when auth is required
662
663## 2.0.22
664
665- Fixed content layout shift when scrolling through slash commands
666- IDE: Add toggle to enable/disable thinking.
667- Fix bug causing duplicate permission prompts with parallel tool calls
668- Add support for enterprise managed MCP allowlist and denylist
669
670## 2.0.21
671
672- Support MCP `structuredContent` field in tool responses
673- Added an interactive question tool
674- Claude will now ask you questions more often in plan mode
675- Added Haiku 4.5 as a model option for Pro users
676- Fixed an issue where queued commands don't have access to previous messages' output
677
678## 2.0.20
679
680- Added support for Claude Skills
681
682## 2.0.19
683
684- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
685- Fixed a bug where Haiku was unnecessarily called in print mode
686
687## 2.0.17
688
689- Added Haiku 4.5 to model selector!
690- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
691- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
692- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
693- OTEL: support HTTP_PROXY and HTTPS_PROXY
694- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
695
696## 2.0.15
697
698- Fixed bug with resuming where previously created files needed to be read again before writing
699- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
700
701## 2.0.14
702
703- Fix @-mentioning MCP servers to toggle them on/off
704- Improve permission checks for bash with inline env vars
705- Fix ultrathink + thinking toggle
706- Reduce unnecessary logins
707- Document --system-prompt
708- Several improvements to rendering
709- Plugins UI polish
710
711## 2.0.13
712
713- Fixed `/plugin` not working on native build
714
715## 2.0.12
716
717- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
718- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
719- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
720- `/plugin validate` command for validating plugin structure and configuration
721- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
722- Plugin documentation available at https://code.claude.com/docs/en/plugins
723- Comprehensive error messages and diagnostics via `/doctor` command
724- Avoid flickering in `/model` selector
725- Improvements to `/help`
726- Avoid mentioning hooks in `/resume` summaries
727- Changes to the "verbose" setting in `/config` now persist across sessions
728
729## 2.0.11
730
731- Reduced system prompt size by 1.4k tokens
732- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
733- Fixed Opus fallback rate limit errors appearing incorrectly
734- Fixed /add-dir command selecting wrong default tab
735
736## 2.0.10
737
738- Rewrote terminal renderer for buttery smooth UI
739- Enable/disable MCP servers by @mentioning, or in /mcp
740- Added tab completion for shell commands in bash mode
741- PreToolUse hooks can now modify tool inputs
742- Press Ctrl-G to edit your prompt in your system's configured text editor
743- Fixes for bash permission checks with environment variables in the command
744
745## 2.0.9
746
747- Fix regression where bash backgrounding stopped working
748
749## 2.0.8
750
751- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
752- IDE: Add drag-and-drop support for files and folders in chat
753- /context: Fix counting for thinking blocks
754- Improve message rendering for users with light themes on dark terminals
755- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
756
757## 2.0.5
758
759- IDE: Fix IME unintended message submission with Enter and Tab
760- IDE: Add "Open in Terminal" link in login screen
761- Fix unhandled OAuth expiration 401 API errors
762- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
763
764## 2.0.1
765
766- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
767- Various bug fixes and presentation improvements
768
769## 2.0.0
770
771- New native VS Code extension
772- Fresh coat of paint throughout the whole app
773- /rewind a conversation to undo code changes
774- /usage command to see plan limits
775- Tab to toggle thinking (sticky across sessions)
776- Ctrl-R to search history
777- Unshipped claude config command
778- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
779- SDK: The Claude Code SDK is now the Claude Agent SDK
780- Add subagents dynamically with `--agents` flag
781
782## 1.0.126
783
784- Enable /context command for Bedrock and Vertex
785- Add mTLS support for HTTP-based OpenTelemetry exporters
786
787## 1.0.124
788
789- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
790- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
791- No longer inform Claude of the list of allowed tools when permission is denied
792- Fixed security vulnerability in Bash tool permission checks
793- Improved VSCode extension performance for large files
794
795## 1.0.123
796
797- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
798- Fixed thinking mode triggering on negation phrases like "don't think"
799- Fixed rendering performance degradation during token streaming
800- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
801- Enhanced BashTool environment snapshot logging
802- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
803- Migrated --debug logging to a file, to enable easy tailing & filtering
804
805## 1.0.120
806
807- Fix input lag during typing, especially noticeable with large prompts
808- Improved VSCode extension command registry and sessions dialog user experience
809- Enhanced sessions dialog responsiveness and visual feedback
810- Fixed IDE compatibility issue by removing worktree support check
811- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
812
813## 1.0.119
814
815- Fix Windows issue where process visually freezes on entering interactive mode
816- Support dynamic headers for MCP servers via headersHelper configuration
817- Fix thinking mode not working in headless sessions
818- Fix slash commands now properly update allowed tools instead of replacing them
819
820## 1.0.117
821
822- Add Ctrl-R history search to recall previous commands like bash/zsh
823- Fix input lag while typing, especially on Windows
824- Add sed command to auto-allowed commands in acceptEdits mode
825- Fix Windows PATH comparison to be case-insensitive for drive letters
826- Add permissions management hint to /add-dir output
827
828## 1.0.115
829
830- Improve thinking mode display with enhanced visual effects
831- Type /t to temporarily disable thinking mode in your prompt
832- Improve path validation for glob and grep tools
833- Show condensed output for post-tool hooks to reduce visual clutter
834- Fix visual feedback when loading state completes
835- Improve UI consistency for permission request dialogs
836
837## 1.0.113
838
839- Deprecated piped input in interactive mode
840- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
841
842## 1.0.112
843
844- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
845- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
846- Hooks: Added systemMessage support for SessionEnd hooks
847- Added `spinnerTipsEnabled` setting to disable spinner tips
848- IDE: Various improvements and bug fixes
849
850## 1.0.111
851
852- /model now validates provided model names
853- Fixed Bash tool crashes caused by malformed shell syntax parsing
854
855## 1.0.110
856
857- /terminal-setup command now supports WezTerm
858- MCP: OAuth tokens now proactively refresh before expiration
859- Fixed reliability issues with background Bash processes
860
861## 1.0.109
862
863- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
864
865## 1.0.106
866
867- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
868
869## 1.0.97
870
871- Settings: /doctor now validates permission rule syntax and suggests corrections
872
873## 1.0.94
874
875- Vertex: add support for global endpoints for supported models
876- /memory command now allows direct editing of all imported memory files
877- SDK: Add custom tools as callbacks
878- Added /todos command to list current todo items
879
880## 1.0.93
881
882- Windows: Add alt + v shortcut for pasting images from clipboard
883- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
884
885## 1.0.90
886
887- Settings file changes take effect immediately - no restart required
888
889## 1.0.88
890
891- Fixed issue causing "OAuth authentication is currently not supported"
892- Status line input now includes `exceeds_200k_tokens`
893- Fixed incorrect usage tracking in /cost.
894- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
895- Bedrock: Updated default Sonnet model to Sonnet 4
896
897## 1.0.86
898
899- Added /context to help users self-serve debug context issues
900- SDK: Added UUID support for all SDK messages
901- SDK: Added `--replay-user-messages` to replay user messages back to stdout
902
903## 1.0.85
904
905- Status line input now includes session cost info
906- Hooks: Introduced SessionEnd hook
907
908## 1.0.84
909
910- Fix tool_use/tool_result id mismatch error when network is unstable
911- Fix Claude sometimes ignoring real-time steering when wrapping up a task
912- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
913- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
914
915## 1.0.83
916
917- @-mention: Support files with spaces in path
918- New shimmering spinner
919
920## 1.0.82
921
922- SDK: Add request cancellation support
923- SDK: New additionalDirectories option to search custom paths, improved slash command processing
924- Settings: Validation prevents invalid fields in .claude/settings.json files
925- MCP: Improve tool name consistency
926- Bash: Fix crash when Claude tries to automatically read large files
927
928## 1.0.81
929
930- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
931- Agents: Fix custom agent loading when agent files are unparsable
932
933## 1.0.80
934
935- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
936
937## 1.0.77
938
939- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
940- SDK: Add session support and permission denial tracking
941- Fix token limit errors in conversation summarization
942- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
943
944## 1.0.73
945
946- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
947- MCP: Press Esc to cancel OAuth authentication flows
948- Bash: Improved command validation and reduced false security warnings
949- UI: Enhanced spinner animations and status line visual hierarchy
950- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
951
952## 1.0.72
953
954- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
955
956## 1.0.71
957
958- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
959- Customizable status line: add your terminal prompt to Claude Code with /statusline
960
961## 1.0.70
962
963- Performance: Optimized message rendering for better performance with large contexts
964- Windows: Fixed native file search, ripgrep, and subagent functionality
965- Added support for @-mentions in slash command arguments
966
967## 1.0.69
968
969- Upgraded Opus to version 4.1
970
971## 1.0.68
972
973- Fix incorrect model names being used for certain commands like `/pr-comments`
974- 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.
975- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
976- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
977- SDK: Added canUseTool callback support for tool confirmation
978- Added `disableAllHooks` setting
979- Improved file suggestions performance in large repos
980
981## 1.0.65
982
983- IDE: Fixed connection stability issues and error handling for diagnostics
984- Windows: Fixed shell environment setup for users without .bashrc files
985
986## 1.0.64
987
988- Agents: Added model customization support - you can now specify which model an agent should use
989- Agents: Fixed unintended access to the recursive agent tool
990- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
991- SDK: Fixed user input tracking across multi-turn conversations
992- Added hidden files to file search and @-mention suggestions
993
994## 1.0.63
995
996- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
997
998## 1.0.62
999
1000- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1001- Hooks: Added SessionStart hook for new session initialization
1002- /add-dir command now supports typeahead for directory paths
1003- Improved network connectivity check reliability
1004
1005## 1.0.61
1006
1007- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1008- Settings: Added `--settings` flag to load settings from a JSON file
1009- Settings: Fixed resolution of settings files paths that are symlinks
1010- OTEL: Fixed reporting of wrong organization after authentication changes
1011- Slash commands: Fixed permissions checking for allowed-tools with Bash
1012- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1013- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1014- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1015
1016## 1.0.60
1017
1018- You can now create custom subagents for specialized tasks! Run /agents to get started
1019
1020## 1.0.59
1021
1022- SDK: Added tool confirmation support with canUseTool callback
1023- SDK: Allow specifying env for spawned process
1024- Hooks: Exposed PermissionDecision to hooks (including "ask")
1025- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1026- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1027
1028## 1.0.58
1029
1030- Added support for reading PDFs
1031- MCP: Improved server health status display in 'claude mcp list'
1032- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1033
1034## 1.0.57
1035
1036- Added support for specifying a model in slash commands
1037- Improved permission messages to help Claude understand allowed tools
1038- Fix: Remove trailing newlines from bash output in terminal wrapping
1039
1040## 1.0.56
1041
1042- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1043- Fixes for WSL IDE detection
1044- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1045
1046## 1.0.55
1047
1048- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1049- Windows: fixed Ctrl+Z crash
1050- SDK: Added ability to capture error logging
1051- Add --system-prompt-file option to override system prompt in print mode
1052
1053## 1.0.54
1054
1055- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1056- Custom slash commands: Added argument-hint to frontmatter
1057- Windows: OAuth uses port 45454 and properly constructs browser URL
1058- Windows: mode switching now uses alt + m, and plan mode renders properly
1059- Shell: Switch to in-memory shell snapshot to fix file-related errors
1060
1061## 1.0.53
1062
1063- Updated @-mention file truncation from 100 lines to 2000 lines
1064- 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).
1065
1066## 1.0.52
1067
1068- Added support for MCP server instructions
1069
1070## 1.0.51
1071
1072- Added support for native Windows (requires Git for Windows)
1073- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1074- Settings: /doctor can now help you identify and fix invalid setting files
1075- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1076- Increased auto-compact warning threshold from 60% to 80%
1077- Fixed an issue with handling user directories with spaces for shell snapshots
1078- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1079- Custom slash commands: Fixed user-level commands in subdirectories
1080- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1081
1082## 1.0.48
1083
1084- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1085- Added progress messages to Bash tool based on the last 5 lines of command output
1086- Added expanding variables support for MCP server configuration
1087- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1088- Improved IDE extension path handling when Claude Code runs in WSL
1089- Hooks: Added a PreCompact hook
1090- Vim mode: Added c, f/F, t/T
1091
1092## 1.0.45
1093
1094- Redesigned Search (Grep) tool with new tool input parameters and features
1095- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1096- Fixed config file corruption issue by enforcing atomic writes
1097- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1098- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1099- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1100
1101## 1.0.44
1102
1103- New /export command lets you quickly export a conversation for sharing
1104- MCP: resource_link tool results are now supported
1105- MCP: tool annotations and tool titles now display in /mcp view
1106- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1107
1108## 1.0.43
1109
1110- Fixed a bug where the theme selector was saving excessively
1111- Hooks: Added EPIPE system error handling
1112
1113## 1.0.42
1114
1115- Added tilde (`~`) expansion support to `/add-dir` command
1116
1117## 1.0.41
1118
1119- Hooks: Split Stop hook triggering into Stop and SubagentStop
1120- Hooks: Enabled optional timeout configuration for each command
1121- Hooks: Added "hook_event_name" to hook input
1122- Fixed a bug where MCP tools would display twice in tool list
1123- New tool parameters JSON for Bash tool in `tool_decision` event
1124
1125## 1.0.40
1126
1127- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1128
1129## 1.0.39
1130
1131- New Active Time metric in OpenTelemetry logging
1132
1133## 1.0.38
1134
1135- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1136
1137## 1.0.37
1138
1139- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1140
1141## 1.0.36
1142
1143- Web search now takes today's date into context
1144- Fixed a bug where stdio MCP servers were not terminating properly on exit
1145
1146## 1.0.35
1147
1148- Added support for MCP OAuth Authorization Server discovery
1149
1150## 1.0.34
1151
1152- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1153
1154## 1.0.33
1155
1156- Improved logging functionality with session ID support
1157- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1158- Improvements to plan mode
1159
1160## 1.0.32
1161
1162- Updated loopback config for litellm
1163- Added forceLoginMethod setting to bypass login selection screen
1164
1165## 1.0.31
1166
1167- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1168
1169## 1.0.30
1170
1171- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1172- Improved file path autocomplete with filename matching
1173- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1174- Enhanced jq regex support for complex filters with pipes and select
1175
1176## 1.0.29
1177
1178- Improved CJK character support in cursor navigation and rendering
1179
1180## 1.0.28
1181
1182- Slash commands: Fix selector display during history navigation
1183- Resizes images before upload to prevent API size limit errors
1184- Added XDG_CONFIG_HOME support to configuration directory
1185- Performance optimizations for memory usage
1186- New attributes (terminal.type, language) in OpenTelemetry logging
1187
1188## 1.0.27
1189
1190- Streamable HTTP MCP servers are now supported
1191- Remote MCP servers (SSE and HTTP) now support OAuth
1192- MCP resources can now be @-mentioned
1193- /resume slash command to switch conversations within Claude Code
1194
1195## 1.0.25
1196
1197- Slash commands: moved "project" and "user" prefixes to descriptions
1198- Slash commands: improved reliability for command discovery
1199- Improved support for Ghostty
1200- Improved web search reliability
1201
1202## 1.0.24
1203
1204- Improved /mcp output
1205- Fixed a bug where settings arrays got overwritten instead of merged
1206
1207## 1.0.23
1208
1209- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1210- Released Python SDK: pip install claude-code-sdk to get started
1211
1212## 1.0.22
1213
1214- SDK: Renamed `total_cost` to `total_cost_usd`
1215
1216## 1.0.21
1217
1218- Improved editing of files with tab-based indentation
1219- Fix for tool_use without matching tool_result errors
1220- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1221
1222## 1.0.18
1223
1224- Added --add-dir CLI argument for specifying additional working directories
1225- Added streaming input support without require -p flag
1226- Improved startup performance and session storage performance
1227- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1228- Added detailed MCP server tools display (/mcp)
1229- MCP authentication and permission improvements
1230- Added auto-reconnection for MCP SSE connections on disconnect
1231- Fixed issue where pasted content was lost when dialogs appeared
1232
1233## 1.0.17
1234
1235- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1236- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1237- MCP server list UI improvements
1238- Update Claude Code process title to display "claude" instead of "node"
1239
1240## 1.0.11
1241
1242- Claude Code can now also be used with a Claude Pro subscription
1243- Added /upgrade for smoother switching to Claude Max plans
1244- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1245- Improved shell configuration error handling
1246- Improved todo list handling during compaction
1247
1248## 1.0.10
1249
1250- Added markdown table support
1251- Improved streaming performance
1252
1253## 1.0.8
1254
1255- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1256- Increased default otel interval from 1s -> 5s
1257- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1258- Fixed a regression where search tools unnecessarily asked for permissions
1259- Added support for triggering thinking non-English languages
1260- Improved compacting UI
1261
1262## 1.0.7
1263
1264- Renamed /allowed-tools -> /permissions
1265- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1266- Deprecated claude config commands in favor of editing settings.json
1267- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1268- Improved error handling for /install-github-app
1269- Bugfixes, UI polish, and tool reliability improvements
1270
1271## 1.0.6
1272
1273- Improved edit reliability for tab-indented files
1274- Respect CLAUDE_CONFIG_DIR everywhere
1275- Reduced unnecessary tool permission prompts
1276- Added support for symlinks in @file typeahead
1277- Bugfixes, UI polish, and tool reliability improvements
1278
1279## 1.0.4
1280
1281- Fixed a bug where MCP tool errors weren't being parsed correctly
1282
1283## 1.0.1
1284
1285- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1286- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1287- Updated documentation links and OAuth process descriptions
1288
1289## 1.0.0
1290
1291- Claude Code is now generally available
1292- Introducing Sonnet 4 and Opus 4 models
1293
1294## 0.2.125
1295
1296- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1297- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1298
1299## 0.2.117
1300
1301- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1302- Introduced settings.cleanupPeriodDays
1303- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1304- Introduced --debug mode
1305
1306## 0.2.108
1307
1308- You can now send messages to Claude while it works to steer Claude in real-time
1309- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1310- Fixed a bug where thinking was not working in -p mode
1311- Fixed a regression in /cost reporting
1312- Deprecated MCP wizard interface in favor of other MCP commands
1313- Lots of other bugfixes and improvements
1314
1315## 0.2.107
1316
1317- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1318
1319## 0.2.106
1320
1321- MCP SSE server configs can now specify custom headers
1322- Fixed a bug where MCP permission prompt didn't always show correctly
1323
1324## 0.2.105
1325
1326- Claude can now search the web
1327- Moved system & account status to /status
1328- Added word movement keybindings for Vim
1329- Improved latency for startup, todo tool, and file edits
1330
1331## 0.2.102
1332
1333- Improved thinking triggering reliability
1334- Improved @mention reliability for images and folders
1335- You can now paste multiple large chunks into one prompt
1336
1337## 0.2.100
1338
1339- Fixed a crash caused by a stack overflow error
1340- Made db storage optional; missing db support disables --continue and --resume
1341
1342## 0.2.98
1343
1344- Fixed an issue where auto-compact was running twice
1345
1346## 0.2.96
1347
1348- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1349
1350## 0.2.93
1351
1352- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1353- Claude now has access to a Todo list that helps it stay on track and be more organized
1354
1355## 0.2.82
1356
1357- Added support for --disallowedTools
1358- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1359
1360## 0.2.75
1361
1362- Hit Enter to queue up additional messages while Claude is working
1363- Drag in or copy/paste image files directly into the prompt
1364- @-mention files to directly add them to context
1365- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1366- Improved performance for filename auto-complete
1367
1368## 0.2.74
1369
1370- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1371- Task tool can now perform writes and run bash commands
1372
1373## 0.2.72
1374
1375- Updated spinner to indicate tokens loaded and tool usage
1376
1377## 0.2.70
1378
1379- Network commands like curl are now available for Claude to use
1380- Claude can now run multiple web queries in parallel
1381- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1382
1383## 0.2.69
1384
1385- Fixed UI glitches with improved Select component behavior
1386- Enhanced terminal output display with better text truncation logic
1387
1388## 0.2.67
1389
1390- Shared project permission rules can be saved in .claude/settings.json
1391
1392## 0.2.66
1393
1394- Print mode (-p) now supports streaming output via --output-format=stream-json
1395- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1396
1397## 0.2.63
1398
1399- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1400
1401## 0.2.61
1402
1403- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1404- Enhanced image detection for more reliable clipboard paste functionality
1405- Fixed an issue where ESC key could crash the conversation history selector
1406
1407## 0.2.59
1408
1409- Copy+paste images directly into your prompt
1410- Improved progress indicators for bash and fetch tools
1411- Bugfixes for non-interactive mode (-p)
1412
1413## 0.2.54
1414
1415- Quickly add to Memory by starting your message with '#'
1416- Press ctrl+r to see full output for long tool results
1417- Added support for MCP SSE transport
1418
1419## 0.2.53
1420
1421- New web fetch tool lets Claude view URLs that you paste in
1422- Fixed a bug with JPEG detection
1423
1424## 0.2.50
1425
1426- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1427
1428## 0.2.49
1429
1430- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1431
1432## 0.2.47
1433
1434- Press Tab to auto-complete file and folder names
1435- Press Shift + Tab to toggle auto-accept for file edits
1436- Automatic conversation compaction for infinite conversation length (toggle with /config)
1437
1438## 0.2.44
1439
1440- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1441
1442## 0.2.41
1443
1444- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1445- MCP server startup no longer blocks the app from starting up
1446
1447## 0.2.37
1448
1449- New /release-notes command lets you view release notes at any time
1450- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1451
1452## 0.2.36
1453
1454- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1455- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1456
1457## 0.2.34
1458
1459- Vim bindings for text input - enable with /vim or /config
1460
1461## 0.2.32
1462
1463- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1464- Fix for some PersistentShell issues
1465
1466## 0.2.31
1467
1468- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1469- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1470
1471## 0.2.30
1472
1473- Added ANSI color theme for better terminal compatibility
1474- Fixed issue where slash command arguments weren't being sent properly
1475- (Mac-only) API keys are now stored in macOS Keychain
1476
1477## 0.2.26
1478
1479- New /approved-tools command for managing tool permissions
1480- Word-level diff display for improved code readability
1481- Fuzzy matching for slash commands
1482
1483## 0.2.21
1484
1485- Fuzzy matching for /commands