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 10e6348 · on Jan 22, 2026 History
1,464 lines · 72.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.17
4
5- Fixed crashes on processors without AVX instruction support
6
7## 2.1.16
8
9- Added new task management system, including new capabilities like dependency tracking
10- [VSCode] Added native plugin management support
11- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
12- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
13- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
14- Fixed session titles on the resume screen not respecting the user's language setting
15- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
16
17## 2.1.15
18
19- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
20- Improved UI rendering performance with React Compiler
21- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
22- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
23
24## 2.1.14
25
26- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
27- Added search to installed plugins list - type to filter by name or description
28- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
29- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
30- Fixed memory issues that could cause crashes when running parallel subagents
31- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
32- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
33- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
34- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
35- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
36- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
37- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
38- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
39- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
40- Improved backspace to delete pasted text as a single token instead of one character at a time
41- [VSCode] Added `/usage` command to display current plan usage
42
43## 2.1.12
44
45- Fixed message rendering bug
46
47## 2.1.11
48
49- Fixed excessive MCP connection requests for HTTP/SSE transports
50
51## 2.1.10
52
53- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
54- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
55- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
56- Improved startup to capture keystrokes typed before the REPL is fully ready
57- Improved file suggestions to show as removable attachments instead of inserting text when accepted
58- [VSCode] Added install count display to plugin listings
59- [VSCode] Added trust warning when installing plugins
60
61## 2.1.9
62
63- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
64- Added `plansDirectory` setting to customize where plan files are stored
65- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
66- Added session URL attribution to commits and PRs created from web sessions
67- Added support for `PreToolUse` hooks to return `additionalContext` to the model
68- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
69- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
70- Fixed MCP server reconnection hanging when cached connection promise never resolves
71- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
72
73## 2.1.7
74
75- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
76- Added ability to provide feedback when accepting permission prompts
77- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
78- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
79- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
80- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
81- 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)
82- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
83- Fixed terminal title animation jitter by using fixed-width braille characters
84- Fixed plugins with git submodules not being fully initialized when installed
85- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
86- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
87- 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.
88- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
89- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
90
91## 2.1.6
92
93- Added search functionality to `/config` command for quickly filtering settings
94- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
95- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
96- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
97- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
98- Added an error display when the editor fails during Ctrl+G
99- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
100- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
101- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
102- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
103- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
104- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
105- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
106- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
107- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
108- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
109- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
110- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
111- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
112- Improved @ autocomplete with icons for different suggestion types and single-line formatting
113- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
114- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
115- Changed terminal title to "Claude Code" on startup for better window identification
116- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
117- [VSCode] Fixed usage indicator not updating after manual compact
118
119## 2.1.5
120
121- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
122
123## 2.1.4
124
125- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
126- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
127
128## 2.1.3
129
130- Merged slash commands and skills, simplifying the mental model with no change in behavior
131- Added release channel (`stable` or `latest`) toggle to `/config`
132- 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
133- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
134- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
135- Fixed mismatch between background task count in status bar and items shown in tasks dialog
136- Fixed sub-agents using the wrong model during conversation compaction
137- Fixed web search in sub-agents using incorrect model
138- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
139- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
140- Improved slash command suggestion readability by truncating long descriptions to 2 lines
141- Changed tool hook execution timeout from 60 seconds to 10 minutes
142- [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)
143
144## 2.1.2
145
146- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
147- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
148- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
149- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
150- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
151- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
152- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
153- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
154- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
155- Fixed updates incorrectly claiming another installation is in progress
156- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
157- Fixed remote session URL and teleport being broken when using `/tasks` command
158- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
159- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
160- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
161- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
162- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
163- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
164- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
165- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
166- [SDK] Changed minimum zod peer dependency to ^4.0.0
167- [VSCode] Fixed usage display not updating after manual compact
168
169## 2.1.0
170
171- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
172- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
173- Added support for `agent` field in skills to specify agent type for execution
174- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
175- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
176- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
177- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
178- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
179- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
180- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
181- Fixed Esc key with queued prompts to only move them to input without canceling the running task
182- Reduced permission prompts for complex bash commands
183- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
184- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
185- Fixed plugin path resolution for file-based marketplace sources
186- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
187- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
188- Fixed Claude in Chrome support for WSL environments
189- Fixed Windows native installer silently failing when executable creation fails
190- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
191- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
192- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
193- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
194- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
195- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
196- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
197- Added hooks support for skill and slash command frontmatter
198- 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
199- Added `/plan` command shortcut to enable plan mode directly from the prompt
200- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
201- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
202- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
203- Added support for `once: true` config for hooks
204- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
205- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
206- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
207- Added left/right arrow key navigation for cycling through tabs in dialogs
208- Added real-time thinking block display in Ctrl+O transcript mode
209- Added filepath to full output in background bash task details dialog
210- Added Skills as a separate category in the context visualization
211- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
212- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
213- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
214- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
215- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
216- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
217- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
218- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
219- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
220- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
221- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
222- Fixed commands with `$()` command substitution failing with parse errors
223- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
224- 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)
225- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
226- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
227- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
228- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
229- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
230- Fixed subagents sometimes not inheriting the parent's model by default
231- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
232- Fixed duplicate Bash commands appearing in permission request option labels
233- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
234- Fixed background task completion notifications to appear proactively with bullet point
235- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
236- Fixed cursor disappearing after dismissing permission dialogs
237- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
238- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
239- Fixed images being silently dropped when queueing messages while backgrounding a task
240- Fixed large pasted images failing with "Image was too large" error
241- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
242- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
243- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
244- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
245- Fixed spinner token counter to properly accumulate tokens from subagents during execution
246- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
247- Fixed race condition where LSP tool could return "no server available" during startup
248- Fixed feedback submission hanging indefinitely when network requests timeout
249- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
250- Fixed hook success message showing trailing colon when hook has no output
251- Multiple optimizations to improve startup performance
252- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
253- Improved performance when reading Jupyter notebooks with many cells
254- Improved reliability for piped input like `cat refactor.md | claude`
255- Improved reliability for AskQuestion tool
256- Improved sed in-place edit commands to render as file edits with diff preview
257- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
258- Improved compaction reliability
259- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
260- Improved skills to show progress while executing, displaying tool uses as they happen
261- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
262- Improved skill suggestions to prioritize recently and frequently used skills
263- Improved spinner feedback when waiting for the first response token
264- Improved token count display in spinner to include tokens from background agents
265- Improved incremental output for async agents to give the main thread more control and visibility
266- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
267- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
268- Improved macOS screenshot paste reliability with TIFF format support
269- Improved `/stats` output
270- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
271- Changed "Interrupted" message color from red to grey for a less alarming appearance
272- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
273- Removed underline styling from image reference links
274- [SDK] Changed minimum zod peer dependency to ^4.0.0
275- [VSCode] Added currently selected model name to the context menu
276- [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")
277- [VSCode] Fixed paragraph breaks not rendering in markdown content
278- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
279- [Windows] Fixed issue with improper rendering
280
281## 2.0.76
282
283- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
284
285## 2.0.75
286
287- Minor bugfixes
288
289## 2.0.74
290
291- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
292- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
293- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
294- Added syntax highlighting info to theme picker
295- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
296- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
297- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
298- Fixed a potential crash when syntax highlighting isn't initialized correctly
299- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
300- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
301- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
302- [Windows] Fixed issue with improper rendering
303- [VSCode] Added gift tag pictogram for year-end promotion message
304
305## 2.0.73
306
307- Added clickable `[Image #N]` links that open attached images in the default viewer
308- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
309- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
310- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
311- Fixed slow input history cycling and race condition that could overwrite text after message submission
312- Improved `/theme` command to open theme picker directly
313- Improved theme picker UI
314- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
315- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
316
317## 2.0.72
318
319- 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
320- Reduced terminal flickering
321- Added scannable QR code to mobile app tip for quick app downloads
322- Added loading indicator when resuming conversations for better feedback
323- Fixed `/context` command not respecting custom system prompts in non-interactive mode
324- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
325- Improved @ mention file suggestion speed (~3x faster in git repositories)
326- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
327- Improved settings validation errors to be more prominent
328- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
329
330## 2.0.71
331
332- Added /config toggle to enable/disable prompt suggestions
333- Added `/settings` as an alias for the `/config` command
334- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
335- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
336- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
337- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
338- New syntax highlighting engine for native build
339
340## 2.0.70
341
342- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
343- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
344- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
345- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
346- Fixed input being cleared when processing queued commands while the user was typing
347- Fixed prompt suggestions replacing typed input when pressing Tab
348- Fixed diff view not updating when terminal is resized
349- Improved memory usage by 3x for large conversations
350- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
351- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
352- Fix thinking mode toggle in /config not persisting correctly
353- Improve UI for file creation permission dialog
354
355## 2.0.69
356
357- Minor bugfixes
358
359## 2.0.68
360
361- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
362- Fixed a bug where disallowed MCP tools were visible to the model
363- Fixed an issue where steering messages could be lost while a subagent is working
364- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
365- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
366- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
367
368## 2.0.67
369
370- Thinking mode is now enabled by default for Opus 4.5
371- Thinking mode configuration has moved to /config
372- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
373- Show reason why autoupdater is disabled in `/doctor`
374- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
375- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
376- Fixed scroll position resetting after deleting a permission rule in `/permissions`
377- 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
378- Fixed `claude install --force` not bypassing stale lock files
379- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
380- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
381
382## 2.0.65
383
384- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
385- Added context window information to status line input
386- Added `fileSuggestion` setting for custom `@` file search commands
387- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
388- Fixed prompt not being saved to history when aborting a query with Escape
389- Fixed Read tool image handling to identify format from bytes instead of file extension
390
391## 2.0.64
392
393- Made auto-compacting instant
394- Agents and bash commands can run asynchronously and send messages to wake up the main agent
395- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
396- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
397- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
398- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
399- Fixed auto-loading .env when using native installer
400- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
401- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
402- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
403- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
404- Bedrock: Improve efficiency of token counting
405- Bedrock: Add support for `aws login` AWS Management Console credentials
406- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
407
408## 2.0.62
409
410- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
411- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
412- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
413- Fixed slash command selection not working when multiple commands share the same name
414- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
415- Fixed running versions getting removed because lock file incorrectly going stale
416- Fixed IDE diff tab not closing when rejecting file changes
417
418## 2.0.61
419
420- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
421
422## 2.0.60
423
424- Added background agent support. Agents run in the background while you work
425- Added --disable-slash-commands CLI flag to disable all slash commands
426- Added model name to "Co-Authored-By" commit messages
427- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
428- Updated Fetch to skip summarization for pre-approved websites
429- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
430
431## 2.0.59
432
433- Added --agent CLI flag to override the agent setting for the current session
434- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
435- VS Code: Fixed .claude.json config file being read from incorrect location
436
437## 2.0.58
438
439- Pro users now have access to Opus 4.5 as part of their subscription!
440- Fixed timer duration showing "11m 60s" instead of "12m 0s"
441- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
442
443## 2.0.57
444
445- Added feedback input when rejecting plans, allowing users to tell Claude what to change
446- VSCode: Added streaming message support for real-time response display
447
448## 2.0.56
449
450- Added setting to enable/disable terminal progress bar (OSC 9;4)
451- 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.
452
453## 2.0.55
454
455- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
456- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
457- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
458- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
459
460## 2.0.54
461
462- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
463- Fix issue with excessive iTerm notifications
464
465## 2.0.52
466
467- Fixed duplicate message display when starting Claude with a command line argument
468- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
469- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
470- Permit some uses of `$!` in bash commands
471
472## 2.0.51
473
474- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
475- Introducing Claude Code for Desktop: https://claude.com/download
476- 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
477- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
478- Plan Mode now builds more precise plans and executes more thoroughly
479- Usage limit notifications now easier to understand
480- Switched `/usage` back to "% used"
481- Fixed handling of thinking errors
482- Fixed performance regression
483
484## 2.0.50
485
486- Fixed bug preventing calling MCP tools that have nested references in their input schemas
487- Silenced a noisy but harmless error during upgrades
488- Improved ultrathink text display
489- Improved clarity of 5-hour session limit warning message
490
491## 2.0.49
492
493- Added readline-style ctrl-y for pasting deleted text
494- Improved clarity of usage limit warning message
495- Fixed handling of subagent permissions
496
497## 2.0.47
498
499- Improved error messages and validation for `claude --teleport`
500- Improved error handling in `/usage`
501- Fixed race condition with history entry not getting logged at exit
502- Fixed Vertex AI configuration not being applied from `settings.json`
503
504## 2.0.46
505
506- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
507
508## 2.0.45
509
510- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
511- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
512- Send background tasks to Claude Code on the web by starting a message with `&`
513
514## 2.0.43
515
516- Added `permissionMode` field for custom agents
517- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
518- Added skills frontmatter field to declare skills to auto-load for subagents
519- Added the `SubagentStart` hook event
520- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
521- Fixed duplicate rendering of some messages in the UI
522- Fixed some visual flickers
523- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
524
525## 2.0.42
526
527- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
528
529## 2.0.41
530
531- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
532- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
533- Fixed incorrect labeling of user settings vs project settings in command descriptions
534- Fixed crash when plugin command hooks timeout during execution
535- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
536- Fixed broken security documentation links in trust dialogs and onboarding
537- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
538- ctrl-r history search landing on a slash command no longer cancels the search
539- SDK: Support custom timeouts for hooks
540- Allow more safe git commands to run without approval
541- Plugins: Added support for sharing and installing output styles
542- Teleporting a session from web will automatically set the upstream branch
543
544## 2.0.37
545
546- Fixed how idleness is computed for notifications
547- Hooks: Added matcher values for Notification hook events
548- Output Styles: Added `keep-coding-instructions` option to frontmatter
549
550## 2.0.36
551
552- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
553- Fixed queued messages being incorrectly executed as bash commands
554- Fixed input being lost when typing while a queued message is processed
555
556## 2.0.35
557
558- Improve fuzzy search results when searching commands
559- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
560- 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
561- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
562- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
563
564## 2.0.34
565
566- VSCode Extension: Added setting to configure the initial permission mode for new conversations
567- Improved file path suggestion performance with native Rust-based fuzzy finder
568- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
569- Fixed memory crash when reading or writing large files (especially base64-encoded images)
570
571## 2.0.33
572
573- Native binary installs now launch quicker.
574- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
575- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
576
577## 2.0.32
578
579- Un-deprecate output styles based on community feedback
580- Added `companyAnnouncements` setting for displaying announcements on startup
581- Fixed hook progress messages not updating correctly during PostToolUse hook execution
582
583## 2.0.31
584
585- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
586- Vertex: add support for Web Search on supported models
587- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
588- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
589- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
590- Fixed plugin uninstall not removing plugins
591
592## 2.0.30
593
594- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
595- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
596- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
597- Added prompt-based stop hooks
598- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
599- Enabled SSE MCP servers on native build
600- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
601- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
602- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
603- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
604- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
605- Fixed bug that saved session permissions to local settings
606- Fixed MCP tools not being available to sub-agents
607- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
608- Fixed delay when navigating through typeahead suggestions with arrow keys
609- VSCode: Restored selection indicator in input footer showing current file or code selection status
610
611## 2.0.28
612
613- Plan mode: introduced new Plan subagent
614- Subagents: claude can now choose to resume subagents
615- Subagents: claude can dynamically choose the model used by its subagents
616- SDK: added --max-budget-usd flag
617- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
618- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
619- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
620- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
621- Various other bug fixes
622
623## 2.0.27
624
625- New UI for permission prompts
626- Added current branch filtering and search to session resume screen for easier navigation
627- Fixed directory @-mention causing "No assistant message found" error
628- VSCode Extension: Add config setting to include .gitignored files in file searches
629- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
630
631## 2.0.25
632
633- 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
634
635## 2.0.24
636
637- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
638- Claude Code Web: Support for Web -> CLI teleport
639- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
640- Bedrock: Display awsAuthRefresh output when auth is required
641
642## 2.0.22
643
644- Fixed content layout shift when scrolling through slash commands
645- IDE: Add toggle to enable/disable thinking.
646- Fix bug causing duplicate permission prompts with parallel tool calls
647- Add support for enterprise managed MCP allowlist and denylist
648
649## 2.0.21
650
651- Support MCP `structuredContent` field in tool responses
652- Added an interactive question tool
653- Claude will now ask you questions more often in plan mode
654- Added Haiku 4.5 as a model option for Pro users
655- Fixed an issue where queued commands don't have access to previous messages' output
656
657## 2.0.20
658
659- Added support for Claude Skills
660
661## 2.0.19
662
663- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
664- Fixed a bug where Haiku was unnecessarily called in print mode
665
666## 2.0.17
667
668- Added Haiku 4.5 to model selector!
669- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
670- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
671- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
672- OTEL: support HTTP_PROXY and HTTPS_PROXY
673- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
674
675## 2.0.15
676
677- Fixed bug with resuming where previously created files needed to be read again before writing
678- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
679
680## 2.0.14
681
682- Fix @-mentioning MCP servers to toggle them on/off
683- Improve permission checks for bash with inline env vars
684- Fix ultrathink + thinking toggle
685- Reduce unnecessary logins
686- Document --system-prompt
687- Several improvements to rendering
688- Plugins UI polish
689
690## 2.0.13
691
692- Fixed `/plugin` not working on native build
693
694## 2.0.12
695
696- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
697- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
698- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
699- `/plugin validate` command for validating plugin structure and configuration
700- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
701- Plugin documentation available at https://code.claude.com/docs/en/plugins
702- Comprehensive error messages and diagnostics via `/doctor` command
703- Avoid flickering in `/model` selector
704- Improvements to `/help`
705- Avoid mentioning hooks in `/resume` summaries
706- Changes to the "verbose" setting in `/config` now persist across sessions
707
708## 2.0.11
709
710- Reduced system prompt size by 1.4k tokens
711- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
712- Fixed Opus fallback rate limit errors appearing incorrectly
713- Fixed /add-dir command selecting wrong default tab
714
715## 2.0.10
716
717- Rewrote terminal renderer for buttery smooth UI
718- Enable/disable MCP servers by @mentioning, or in /mcp
719- Added tab completion for shell commands in bash mode
720- PreToolUse hooks can now modify tool inputs
721- Press Ctrl-G to edit your prompt in your system's configured text editor
722- Fixes for bash permission checks with environment variables in the command
723
724## 2.0.9
725
726- Fix regression where bash backgrounding stopped working
727
728## 2.0.8
729
730- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
731- IDE: Add drag-and-drop support for files and folders in chat
732- /context: Fix counting for thinking blocks
733- Improve message rendering for users with light themes on dark terminals
734- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
735
736## 2.0.5
737
738- IDE: Fix IME unintended message submission with Enter and Tab
739- IDE: Add "Open in Terminal" link in login screen
740- Fix unhandled OAuth expiration 401 API errors
741- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
742
743## 2.0.1
744
745- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
746- Various bug fixes and presentation improvements
747
748## 2.0.0
749
750- New native VS Code extension
751- Fresh coat of paint throughout the whole app
752- /rewind a conversation to undo code changes
753- /usage command to see plan limits
754- Tab to toggle thinking (sticky across sessions)
755- Ctrl-R to search history
756- Unshipped claude config command
757- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
758- SDK: The Claude Code SDK is now the Claude Agent SDK
759- Add subagents dynamically with `--agents` flag
760
761## 1.0.126
762
763- Enable /context command for Bedrock and Vertex
764- Add mTLS support for HTTP-based OpenTelemetry exporters
765
766## 1.0.124
767
768- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
769- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
770- No longer inform Claude of the list of allowed tools when permission is denied
771- Fixed security vulnerability in Bash tool permission checks
772- Improved VSCode extension performance for large files
773
774## 1.0.123
775
776- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
777- Fixed thinking mode triggering on negation phrases like "don't think"
778- Fixed rendering performance degradation during token streaming
779- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
780- Enhanced BashTool environment snapshot logging
781- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
782- Migrated --debug logging to a file, to enable easy tailing & filtering
783
784## 1.0.120
785
786- Fix input lag during typing, especially noticeable with large prompts
787- Improved VSCode extension command registry and sessions dialog user experience
788- Enhanced sessions dialog responsiveness and visual feedback
789- Fixed IDE compatibility issue by removing worktree support check
790- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
791
792## 1.0.119
793
794- Fix Windows issue where process visually freezes on entering interactive mode
795- Support dynamic headers for MCP servers via headersHelper configuration
796- Fix thinking mode not working in headless sessions
797- Fix slash commands now properly update allowed tools instead of replacing them
798
799## 1.0.117
800
801- Add Ctrl-R history search to recall previous commands like bash/zsh
802- Fix input lag while typing, especially on Windows
803- Add sed command to auto-allowed commands in acceptEdits mode
804- Fix Windows PATH comparison to be case-insensitive for drive letters
805- Add permissions management hint to /add-dir output
806
807## 1.0.115
808
809- Improve thinking mode display with enhanced visual effects
810- Type /t to temporarily disable thinking mode in your prompt
811- Improve path validation for glob and grep tools
812- Show condensed output for post-tool hooks to reduce visual clutter
813- Fix visual feedback when loading state completes
814- Improve UI consistency for permission request dialogs
815
816## 1.0.113
817
818- Deprecated piped input in interactive mode
819- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
820
821## 1.0.112
822
823- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
824- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
825- Hooks: Added systemMessage support for SessionEnd hooks
826- Added `spinnerTipsEnabled` setting to disable spinner tips
827- IDE: Various improvements and bug fixes
828
829## 1.0.111
830
831- /model now validates provided model names
832- Fixed Bash tool crashes caused by malformed shell syntax parsing
833
834## 1.0.110
835
836- /terminal-setup command now supports WezTerm
837- MCP: OAuth tokens now proactively refresh before expiration
838- Fixed reliability issues with background Bash processes
839
840## 1.0.109
841
842- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
843
844## 1.0.106
845
846- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
847
848## 1.0.97
849
850- Settings: /doctor now validates permission rule syntax and suggests corrections
851
852## 1.0.94
853
854- Vertex: add support for global endpoints for supported models
855- /memory command now allows direct editing of all imported memory files
856- SDK: Add custom tools as callbacks
857- Added /todos command to list current todo items
858
859## 1.0.93
860
861- Windows: Add alt + v shortcut for pasting images from clipboard
862- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
863
864## 1.0.90
865
866- Settings file changes take effect immediately - no restart required
867
868## 1.0.88
869
870- Fixed issue causing "OAuth authentication is currently not supported"
871- Status line input now includes `exceeds_200k_tokens`
872- Fixed incorrect usage tracking in /cost.
873- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
874- Bedrock: Updated default Sonnet model to Sonnet 4
875
876## 1.0.86
877
878- Added /context to help users self-serve debug context issues
879- SDK: Added UUID support for all SDK messages
880- SDK: Added `--replay-user-messages` to replay user messages back to stdout
881
882## 1.0.85
883
884- Status line input now includes session cost info
885- Hooks: Introduced SessionEnd hook
886
887## 1.0.84
888
889- Fix tool_use/tool_result id mismatch error when network is unstable
890- Fix Claude sometimes ignoring real-time steering when wrapping up a task
891- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
892- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
893
894## 1.0.83
895
896- @-mention: Support files with spaces in path
897- New shimmering spinner
898
899## 1.0.82
900
901- SDK: Add request cancellation support
902- SDK: New additionalDirectories option to search custom paths, improved slash command processing
903- Settings: Validation prevents invalid fields in .claude/settings.json files
904- MCP: Improve tool name consistency
905- Bash: Fix crash when Claude tries to automatically read large files
906
907## 1.0.81
908
909- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
910- Agents: Fix custom agent loading when agent files are unparsable
911
912## 1.0.80
913
914- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
915
916## 1.0.77
917
918- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
919- SDK: Add session support and permission denial tracking
920- Fix token limit errors in conversation summarization
921- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
922
923## 1.0.73
924
925- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
926- MCP: Press Esc to cancel OAuth authentication flows
927- Bash: Improved command validation and reduced false security warnings
928- UI: Enhanced spinner animations and status line visual hierarchy
929- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
930
931## 1.0.72
932
933- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
934
935## 1.0.71
936
937- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
938- Customizable status line: add your terminal prompt to Claude Code with /statusline
939
940## 1.0.70
941
942- Performance: Optimized message rendering for better performance with large contexts
943- Windows: Fixed native file search, ripgrep, and subagent functionality
944- Added support for @-mentions in slash command arguments
945
946## 1.0.69
947
948- Upgraded Opus to version 4.1
949
950## 1.0.68
951
952- Fix incorrect model names being used for certain commands like `/pr-comments`
953- 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.
954- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
955- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
956- SDK: Added canUseTool callback support for tool confirmation
957- Added `disableAllHooks` setting
958- Improved file suggestions performance in large repos
959
960## 1.0.65
961
962- IDE: Fixed connection stability issues and error handling for diagnostics
963- Windows: Fixed shell environment setup for users without .bashrc files
964
965## 1.0.64
966
967- Agents: Added model customization support - you can now specify which model an agent should use
968- Agents: Fixed unintended access to the recursive agent tool
969- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
970- SDK: Fixed user input tracking across multi-turn conversations
971- Added hidden files to file search and @-mention suggestions
972
973## 1.0.63
974
975- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
976
977## 1.0.62
978
979- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
980- Hooks: Added SessionStart hook for new session initialization
981- /add-dir command now supports typeahead for directory paths
982- Improved network connectivity check reliability
983
984## 1.0.61
985
986- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
987- Settings: Added `--settings` flag to load settings from a JSON file
988- Settings: Fixed resolution of settings files paths that are symlinks
989- OTEL: Fixed reporting of wrong organization after authentication changes
990- Slash commands: Fixed permissions checking for allowed-tools with Bash
991- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
992- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
993- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
994
995## 1.0.60
996
997- You can now create custom subagents for specialized tasks! Run /agents to get started
998
999## 1.0.59
1000
1001- SDK: Added tool confirmation support with canUseTool callback
1002- SDK: Allow specifying env for spawned process
1003- Hooks: Exposed PermissionDecision to hooks (including "ask")
1004- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1005- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1006
1007## 1.0.58
1008
1009- Added support for reading PDFs
1010- MCP: Improved server health status display in 'claude mcp list'
1011- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1012
1013## 1.0.57
1014
1015- Added support for specifying a model in slash commands
1016- Improved permission messages to help Claude understand allowed tools
1017- Fix: Remove trailing newlines from bash output in terminal wrapping
1018
1019## 1.0.56
1020
1021- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1022- Fixes for WSL IDE detection
1023- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1024
1025## 1.0.55
1026
1027- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1028- Windows: fixed Ctrl+Z crash
1029- SDK: Added ability to capture error logging
1030- Add --system-prompt-file option to override system prompt in print mode
1031
1032## 1.0.54
1033
1034- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1035- Custom slash commands: Added argument-hint to frontmatter
1036- Windows: OAuth uses port 45454 and properly constructs browser URL
1037- Windows: mode switching now uses alt + m, and plan mode renders properly
1038- Shell: Switch to in-memory shell snapshot to fix file-related errors
1039
1040## 1.0.53
1041
1042- Updated @-mention file truncation from 100 lines to 2000 lines
1043- 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).
1044
1045## 1.0.52
1046
1047- Added support for MCP server instructions
1048
1049## 1.0.51
1050
1051- Added support for native Windows (requires Git for Windows)
1052- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1053- Settings: /doctor can now help you identify and fix invalid setting files
1054- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1055- Increased auto-compact warning threshold from 60% to 80%
1056- Fixed an issue with handling user directories with spaces for shell snapshots
1057- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1058- Custom slash commands: Fixed user-level commands in subdirectories
1059- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1060
1061## 1.0.48
1062
1063- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1064- Added progress messages to Bash tool based on the last 5 lines of command output
1065- Added expanding variables support for MCP server configuration
1066- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1067- Improved IDE extension path handling when Claude Code runs in WSL
1068- Hooks: Added a PreCompact hook
1069- Vim mode: Added c, f/F, t/T
1070
1071## 1.0.45
1072
1073- Redesigned Search (Grep) tool with new tool input parameters and features
1074- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1075- Fixed config file corruption issue by enforcing atomic writes
1076- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1077- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1078- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1079
1080## 1.0.44
1081
1082- New /export command lets you quickly export a conversation for sharing
1083- MCP: resource_link tool results are now supported
1084- MCP: tool annotations and tool titles now display in /mcp view
1085- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1086
1087## 1.0.43
1088
1089- Fixed a bug where the theme selector was saving excessively
1090- Hooks: Added EPIPE system error handling
1091
1092## 1.0.42
1093
1094- Added tilde (`~`) expansion support to `/add-dir` command
1095
1096## 1.0.41
1097
1098- Hooks: Split Stop hook triggering into Stop and SubagentStop
1099- Hooks: Enabled optional timeout configuration for each command
1100- Hooks: Added "hook_event_name" to hook input
1101- Fixed a bug where MCP tools would display twice in tool list
1102- New tool parameters JSON for Bash tool in `tool_decision` event
1103
1104## 1.0.40
1105
1106- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1107
1108## 1.0.39
1109
1110- New Active Time metric in OpenTelemetry logging
1111
1112## 1.0.38
1113
1114- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1115
1116## 1.0.37
1117
1118- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1119
1120## 1.0.36
1121
1122- Web search now takes today's date into context
1123- Fixed a bug where stdio MCP servers were not terminating properly on exit
1124
1125## 1.0.35
1126
1127- Added support for MCP OAuth Authorization Server discovery
1128
1129## 1.0.34
1130
1131- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1132
1133## 1.0.33
1134
1135- Improved logging functionality with session ID support
1136- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1137- Improvements to plan mode
1138
1139## 1.0.32
1140
1141- Updated loopback config for litellm
1142- Added forceLoginMethod setting to bypass login selection screen
1143
1144## 1.0.31
1145
1146- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1147
1148## 1.0.30
1149
1150- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1151- Improved file path autocomplete with filename matching
1152- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1153- Enhanced jq regex support for complex filters with pipes and select
1154
1155## 1.0.29
1156
1157- Improved CJK character support in cursor navigation and rendering
1158
1159## 1.0.28
1160
1161- Slash commands: Fix selector display during history navigation
1162- Resizes images before upload to prevent API size limit errors
1163- Added XDG_CONFIG_HOME support to configuration directory
1164- Performance optimizations for memory usage
1165- New attributes (terminal.type, language) in OpenTelemetry logging
1166
1167## 1.0.27
1168
1169- Streamable HTTP MCP servers are now supported
1170- Remote MCP servers (SSE and HTTP) now support OAuth
1171- MCP resources can now be @-mentioned
1172- /resume slash command to switch conversations within Claude Code
1173
1174## 1.0.25
1175
1176- Slash commands: moved "project" and "user" prefixes to descriptions
1177- Slash commands: improved reliability for command discovery
1178- Improved support for Ghostty
1179- Improved web search reliability
1180
1181## 1.0.24
1182
1183- Improved /mcp output
1184- Fixed a bug where settings arrays got overwritten instead of merged
1185
1186## 1.0.23
1187
1188- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1189- Released Python SDK: pip install claude-code-sdk to get started
1190
1191## 1.0.22
1192
1193- SDK: Renamed `total_cost` to `total_cost_usd`
1194
1195## 1.0.21
1196
1197- Improved editing of files with tab-based indentation
1198- Fix for tool_use without matching tool_result errors
1199- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1200
1201## 1.0.18
1202
1203- Added --add-dir CLI argument for specifying additional working directories
1204- Added streaming input support without require -p flag
1205- Improved startup performance and session storage performance
1206- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1207- Added detailed MCP server tools display (/mcp)
1208- MCP authentication and permission improvements
1209- Added auto-reconnection for MCP SSE connections on disconnect
1210- Fixed issue where pasted content was lost when dialogs appeared
1211
1212## 1.0.17
1213
1214- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1215- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1216- MCP server list UI improvements
1217- Update Claude Code process title to display "claude" instead of "node"
1218
1219## 1.0.11
1220
1221- Claude Code can now also be used with a Claude Pro subscription
1222- Added /upgrade for smoother switching to Claude Max plans
1223- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1224- Improved shell configuration error handling
1225- Improved todo list handling during compaction
1226
1227## 1.0.10
1228
1229- Added markdown table support
1230- Improved streaming performance
1231
1232## 1.0.8
1233
1234- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1235- Increased default otel interval from 1s -> 5s
1236- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1237- Fixed a regression where search tools unnecessarily asked for permissions
1238- Added support for triggering thinking non-English languages
1239- Improved compacting UI
1240
1241## 1.0.7
1242
1243- Renamed /allowed-tools -> /permissions
1244- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1245- Deprecated claude config commands in favor of editing settings.json
1246- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1247- Improved error handling for /install-github-app
1248- Bugfixes, UI polish, and tool reliability improvements
1249
1250## 1.0.6
1251
1252- Improved edit reliability for tab-indented files
1253- Respect CLAUDE_CONFIG_DIR everywhere
1254- Reduced unnecessary tool permission prompts
1255- Added support for symlinks in @file typeahead
1256- Bugfixes, UI polish, and tool reliability improvements
1257
1258## 1.0.4
1259
1260- Fixed a bug where MCP tool errors weren't being parsed correctly
1261
1262## 1.0.1
1263
1264- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1265- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1266- Updated documentation links and OAuth process descriptions
1267
1268## 1.0.0
1269
1270- Claude Code is now generally available
1271- Introducing Sonnet 4 and Opus 4 models
1272
1273## 0.2.125
1274
1275- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1276- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1277
1278## 0.2.117
1279
1280- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1281- Introduced settings.cleanupPeriodDays
1282- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1283- Introduced --debug mode
1284
1285## 0.2.108
1286
1287- You can now send messages to Claude while it works to steer Claude in real-time
1288- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1289- Fixed a bug where thinking was not working in -p mode
1290- Fixed a regression in /cost reporting
1291- Deprecated MCP wizard interface in favor of other MCP commands
1292- Lots of other bugfixes and improvements
1293
1294## 0.2.107
1295
1296- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1297
1298## 0.2.106
1299
1300- MCP SSE server configs can now specify custom headers
1301- Fixed a bug where MCP permission prompt didn't always show correctly
1302
1303## 0.2.105
1304
1305- Claude can now search the web
1306- Moved system & account status to /status
1307- Added word movement keybindings for Vim
1308- Improved latency for startup, todo tool, and file edits
1309
1310## 0.2.102
1311
1312- Improved thinking triggering reliability
1313- Improved @mention reliability for images and folders
1314- You can now paste multiple large chunks into one prompt
1315
1316## 0.2.100
1317
1318- Fixed a crash caused by a stack overflow error
1319- Made db storage optional; missing db support disables --continue and --resume
1320
1321## 0.2.98
1322
1323- Fixed an issue where auto-compact was running twice
1324
1325## 0.2.96
1326
1327- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1328
1329## 0.2.93
1330
1331- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1332- Claude now has access to a Todo list that helps it stay on track and be more organized
1333
1334## 0.2.82
1335
1336- Added support for --disallowedTools
1337- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1338
1339## 0.2.75
1340
1341- Hit Enter to queue up additional messages while Claude is working
1342- Drag in or copy/paste image files directly into the prompt
1343- @-mention files to directly add them to context
1344- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1345- Improved performance for filename auto-complete
1346
1347## 0.2.74
1348
1349- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1350- Task tool can now perform writes and run bash commands
1351
1352## 0.2.72
1353
1354- Updated spinner to indicate tokens loaded and tool usage
1355
1356## 0.2.70
1357
1358- Network commands like curl are now available for Claude to use
1359- Claude can now run multiple web queries in parallel
1360- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1361
1362## 0.2.69
1363
1364- Fixed UI glitches with improved Select component behavior
1365- Enhanced terminal output display with better text truncation logic
1366
1367## 0.2.67
1368
1369- Shared project permission rules can be saved in .claude/settings.json
1370
1371## 0.2.66
1372
1373- Print mode (-p) now supports streaming output via --output-format=stream-json
1374- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1375
1376## 0.2.63
1377
1378- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1379
1380## 0.2.61
1381
1382- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1383- Enhanced image detection for more reliable clipboard paste functionality
1384- Fixed an issue where ESC key could crash the conversation history selector
1385
1386## 0.2.59
1387
1388- Copy+paste images directly into your prompt
1389- Improved progress indicators for bash and fetch tools
1390- Bugfixes for non-interactive mode (-p)
1391
1392## 0.2.54
1393
1394- Quickly add to Memory by starting your message with '#'
1395- Press ctrl+r to see full output for long tool results
1396- Added support for MCP SSE transport
1397
1398## 0.2.53
1399
1400- New web fetch tool lets Claude view URLs that you paste in
1401- Fixed a bug with JPEG detection
1402
1403## 0.2.50
1404
1405- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1406
1407## 0.2.49
1408
1409- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1410
1411## 0.2.47
1412
1413- Press Tab to auto-complete file and folder names
1414- Press Shift + Tab to toggle auto-accept for file edits
1415- Automatic conversation compaction for infinite conversation length (toggle with /config)
1416
1417## 0.2.44
1418
1419- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1420
1421## 0.2.41
1422
1423- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1424- MCP server startup no longer blocks the app from starting up
1425
1426## 0.2.37
1427
1428- New /release-notes command lets you view release notes at any time
1429- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1430
1431## 0.2.36
1432
1433- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1434- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1435
1436## 0.2.34
1437
1438- Vim bindings for text input - enable with /vim or /config
1439
1440## 0.2.32
1441
1442- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1443- Fix for some PersistentShell issues
1444
1445## 0.2.31
1446
1447- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1448- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1449
1450## 0.2.30
1451
1452- Added ANSI color theme for better terminal compatibility
1453- Fixed issue where slash command arguments weren't being sent properly
1454- (Mac-only) API keys are now stored in macOS Keychain
1455
1456## 0.2.26
1457
1458- New /approved-tools command for managing tool permissions
1459- Word-level diff display for improved code readability
1460- Fuzzy matching for slash commands
1461
1462## 0.2.21
1463
1464- Fuzzy matching for /commands