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 e431f5b · on Jan 22, 2026 History
1,460 lines · 72.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.16
4
5- Added new task management system, including new capabilities like dependency tracking
6- [VSCode] Added native plugin management support
7- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
8- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
9- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
10- Fixed session titles on the resume screen not respecting the user's language setting
11- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
12
13## 2.1.15
14
15- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
16- Improved UI rendering performance with React Compiler
17- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
18- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
19
20## 2.1.14
21
22- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
23- Added search to installed plugins list - type to filter by name or description
24- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
25- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
26- Fixed memory issues that could cause crashes when running parallel subagents
27- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
28- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
29- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
30- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
31- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
32- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
33- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
34- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
35- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
36- Improved backspace to delete pasted text as a single token instead of one character at a time
37- [VSCode] Added `/usage` command to display current plan usage
38
39## 2.1.12
40
41- Fixed message rendering bug
42
43## 2.1.11
44
45- Fixed excessive MCP connection requests for HTTP/SSE transports
46
47## 2.1.10
48
49- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
50- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
51- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
52- Improved startup to capture keystrokes typed before the REPL is fully ready
53- Improved file suggestions to show as removable attachments instead of inserting text when accepted
54- [VSCode] Added install count display to plugin listings
55- [VSCode] Added trust warning when installing plugins
56
57## 2.1.9
58
59- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
60- Added `plansDirectory` setting to customize where plan files are stored
61- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
62- Added session URL attribution to commits and PRs created from web sessions
63- Added support for `PreToolUse` hooks to return `additionalContext` to the model
64- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
65- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
66- Fixed MCP server reconnection hanging when cached connection promise never resolves
67- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
68
69## 2.1.7
70
71- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
72- Added ability to provide feedback when accepting permission prompts
73- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
74- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
75- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
76- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
77- 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)
78- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
79- Fixed terminal title animation jitter by using fixed-width braille characters
80- Fixed plugins with git submodules not being fully initialized when installed
81- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
82- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
83- 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.
84- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
85- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
86
87## 2.1.6
88
89- Added search functionality to `/config` command for quickly filtering settings
90- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
91- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
92- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
93- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
94- Added an error display when the editor fails during Ctrl+G
95- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
96- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
97- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
98- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
99- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
100- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
101- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
102- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
103- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
104- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
105- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
106- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
107- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
108- Improved @ autocomplete with icons for different suggestion types and single-line formatting
109- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
110- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
111- Changed terminal title to "Claude Code" on startup for better window identification
112- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
113- [VSCode] Fixed usage indicator not updating after manual compact
114
115## 2.1.5
116
117- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
118
119## 2.1.4
120
121- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
122- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
123
124## 2.1.3
125
126- Merged slash commands and skills, simplifying the mental model with no change in behavior
127- Added release channel (`stable` or `latest`) toggle to `/config`
128- 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
129- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
130- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
131- Fixed mismatch between background task count in status bar and items shown in tasks dialog
132- Fixed sub-agents using the wrong model during conversation compaction
133- Fixed web search in sub-agents using incorrect model
134- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
135- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
136- Improved slash command suggestion readability by truncating long descriptions to 2 lines
137- Changed tool hook execution timeout from 60 seconds to 10 minutes
138- [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)
139
140## 2.1.2
141
142- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
143- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
144- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
145- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
146- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
147- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
148- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
149- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
150- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
151- Fixed updates incorrectly claiming another installation is in progress
152- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
153- Fixed remote session URL and teleport being broken when using `/tasks` command
154- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
155- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
156- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
157- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
158- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
159- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
160- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
161- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
162- [SDK] Changed minimum zod peer dependency to ^4.0.0
163- [VSCode] Fixed usage display not updating after manual compact
164
165## 2.1.0
166
167- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
168- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
169- Added support for `agent` field in skills to specify agent type for execution
170- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
171- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
172- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
173- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
174- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
175- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
176- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
177- Fixed Esc key with queued prompts to only move them to input without canceling the running task
178- Reduced permission prompts for complex bash commands
179- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
180- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
181- Fixed plugin path resolution for file-based marketplace sources
182- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
183- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
184- Fixed Claude in Chrome support for WSL environments
185- Fixed Windows native installer silently failing when executable creation fails
186- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
187- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
188- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
189- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
190- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
191- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
192- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
193- Added hooks support for skill and slash command frontmatter
194- 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
195- Added `/plan` command shortcut to enable plan mode directly from the prompt
196- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
197- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
198- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
199- Added support for `once: true` config for hooks
200- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
201- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
202- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
203- Added left/right arrow key navigation for cycling through tabs in dialogs
204- Added real-time thinking block display in Ctrl+O transcript mode
205- Added filepath to full output in background bash task details dialog
206- Added Skills as a separate category in the context visualization
207- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
208- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
209- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
210- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
211- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
212- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
213- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
214- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
215- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
216- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
217- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
218- Fixed commands with `$()` command substitution failing with parse errors
219- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
220- 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)
221- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
222- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
223- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
224- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
225- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
226- Fixed subagents sometimes not inheriting the parent's model by default
227- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
228- Fixed duplicate Bash commands appearing in permission request option labels
229- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
230- Fixed background task completion notifications to appear proactively with bullet point
231- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
232- Fixed cursor disappearing after dismissing permission dialogs
233- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
234- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
235- Fixed images being silently dropped when queueing messages while backgrounding a task
236- Fixed large pasted images failing with "Image was too large" error
237- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
238- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
239- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
240- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
241- Fixed spinner token counter to properly accumulate tokens from subagents during execution
242- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
243- Fixed race condition where LSP tool could return "no server available" during startup
244- Fixed feedback submission hanging indefinitely when network requests timeout
245- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
246- Fixed hook success message showing trailing colon when hook has no output
247- Multiple optimizations to improve startup performance
248- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
249- Improved performance when reading Jupyter notebooks with many cells
250- Improved reliability for piped input like `cat refactor.md | claude`
251- Improved reliability for AskQuestion tool
252- Improved sed in-place edit commands to render as file edits with diff preview
253- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
254- Improved compaction reliability
255- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
256- Improved skills to show progress while executing, displaying tool uses as they happen
257- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
258- Improved skill suggestions to prioritize recently and frequently used skills
259- Improved spinner feedback when waiting for the first response token
260- Improved token count display in spinner to include tokens from background agents
261- Improved incremental output for async agents to give the main thread more control and visibility
262- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
263- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
264- Improved macOS screenshot paste reliability with TIFF format support
265- Improved `/stats` output
266- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
267- Changed "Interrupted" message color from red to grey for a less alarming appearance
268- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
269- Removed underline styling from image reference links
270- [SDK] Changed minimum zod peer dependency to ^4.0.0
271- [VSCode] Added currently selected model name to the context menu
272- [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")
273- [VSCode] Fixed paragraph breaks not rendering in markdown content
274- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
275- [Windows] Fixed issue with improper rendering
276
277## 2.0.76
278
279- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
280
281## 2.0.75
282
283- Minor bugfixes
284
285## 2.0.74
286
287- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
288- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
289- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
290- Added syntax highlighting info to theme picker
291- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
292- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
293- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
294- Fixed a potential crash when syntax highlighting isn't initialized correctly
295- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
296- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
297- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
298- [Windows] Fixed issue with improper rendering
299- [VSCode] Added gift tag pictogram for year-end promotion message
300
301## 2.0.73
302
303- Added clickable `[Image #N]` links that open attached images in the default viewer
304- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
305- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
306- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
307- Fixed slow input history cycling and race condition that could overwrite text after message submission
308- Improved `/theme` command to open theme picker directly
309- Improved theme picker UI
310- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
311- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
312
313## 2.0.72
314
315- 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
316- Reduced terminal flickering
317- Added scannable QR code to mobile app tip for quick app downloads
318- Added loading indicator when resuming conversations for better feedback
319- Fixed `/context` command not respecting custom system prompts in non-interactive mode
320- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
321- Improved @ mention file suggestion speed (~3x faster in git repositories)
322- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
323- Improved settings validation errors to be more prominent
324- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
325
326## 2.0.71
327
328- Added /config toggle to enable/disable prompt suggestions
329- Added `/settings` as an alias for the `/config` command
330- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
331- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
332- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
333- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
334- New syntax highlighting engine for native build
335
336## 2.0.70
337
338- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
339- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
340- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
341- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
342- Fixed input being cleared when processing queued commands while the user was typing
343- Fixed prompt suggestions replacing typed input when pressing Tab
344- Fixed diff view not updating when terminal is resized
345- Improved memory usage by 3x for large conversations
346- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
347- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
348- Fix thinking mode toggle in /config not persisting correctly
349- Improve UI for file creation permission dialog
350
351## 2.0.69
352
353- Minor bugfixes
354
355## 2.0.68
356
357- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
358- Fixed a bug where disallowed MCP tools were visible to the model
359- Fixed an issue where steering messages could be lost while a subagent is working
360- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
361- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
362- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
363
364## 2.0.67
365
366- Thinking mode is now enabled by default for Opus 4.5
367- Thinking mode configuration has moved to /config
368- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
369- Show reason why autoupdater is disabled in `/doctor`
370- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
371- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
372- Fixed scroll position resetting after deleting a permission rule in `/permissions`
373- 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
374- Fixed `claude install --force` not bypassing stale lock files
375- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
376- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
377
378## 2.0.65
379
380- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
381- Added context window information to status line input
382- Added `fileSuggestion` setting for custom `@` file search commands
383- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
384- Fixed prompt not being saved to history when aborting a query with Escape
385- Fixed Read tool image handling to identify format from bytes instead of file extension
386
387## 2.0.64
388
389- Made auto-compacting instant
390- Agents and bash commands can run asynchronously and send messages to wake up the main agent
391- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
392- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
393- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
394- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
395- Fixed auto-loading .env when using native installer
396- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
397- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
398- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
399- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
400- Bedrock: Improve efficiency of token counting
401- Bedrock: Add support for `aws login` AWS Management Console credentials
402- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
403
404## 2.0.62
405
406- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
407- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
408- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
409- Fixed slash command selection not working when multiple commands share the same name
410- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
411- Fixed running versions getting removed because lock file incorrectly going stale
412- Fixed IDE diff tab not closing when rejecting file changes
413
414## 2.0.61
415
416- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
417
418## 2.0.60
419
420- Added background agent support. Agents run in the background while you work
421- Added --disable-slash-commands CLI flag to disable all slash commands
422- Added model name to "Co-Authored-By" commit messages
423- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
424- Updated Fetch to skip summarization for pre-approved websites
425- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
426
427## 2.0.59
428
429- Added --agent CLI flag to override the agent setting for the current session
430- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
431- VS Code: Fixed .claude.json config file being read from incorrect location
432
433## 2.0.58
434
435- Pro users now have access to Opus 4.5 as part of their subscription!
436- Fixed timer duration showing "11m 60s" instead of "12m 0s"
437- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
438
439## 2.0.57
440
441- Added feedback input when rejecting plans, allowing users to tell Claude what to change
442- VSCode: Added streaming message support for real-time response display
443
444## 2.0.56
445
446- Added setting to enable/disable terminal progress bar (OSC 9;4)
447- 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.
448
449## 2.0.55
450
451- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
452- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
453- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
454- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
455
456## 2.0.54
457
458- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
459- Fix issue with excessive iTerm notifications
460
461## 2.0.52
462
463- Fixed duplicate message display when starting Claude with a command line argument
464- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
465- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
466- Permit some uses of `$!` in bash commands
467
468## 2.0.51
469
470- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
471- Introducing Claude Code for Desktop: https://claude.com/download
472- 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
473- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
474- Plan Mode now builds more precise plans and executes more thoroughly
475- Usage limit notifications now easier to understand
476- Switched `/usage` back to "% used"
477- Fixed handling of thinking errors
478- Fixed performance regression
479
480## 2.0.50
481
482- Fixed bug preventing calling MCP tools that have nested references in their input schemas
483- Silenced a noisy but harmless error during upgrades
484- Improved ultrathink text display
485- Improved clarity of 5-hour session limit warning message
486
487## 2.0.49
488
489- Added readline-style ctrl-y for pasting deleted text
490- Improved clarity of usage limit warning message
491- Fixed handling of subagent permissions
492
493## 2.0.47
494
495- Improved error messages and validation for `claude --teleport`
496- Improved error handling in `/usage`
497- Fixed race condition with history entry not getting logged at exit
498- Fixed Vertex AI configuration not being applied from `settings.json`
499
500## 2.0.46
501
502- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
503
504## 2.0.45
505
506- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
507- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
508- Send background tasks to Claude Code on the web by starting a message with `&`
509
510## 2.0.43
511
512- Added `permissionMode` field for custom agents
513- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
514- Added skills frontmatter field to declare skills to auto-load for subagents
515- Added the `SubagentStart` hook event
516- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
517- Fixed duplicate rendering of some messages in the UI
518- Fixed some visual flickers
519- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
520
521## 2.0.42
522
523- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
524
525## 2.0.41
526
527- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
528- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
529- Fixed incorrect labeling of user settings vs project settings in command descriptions
530- Fixed crash when plugin command hooks timeout during execution
531- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
532- Fixed broken security documentation links in trust dialogs and onboarding
533- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
534- ctrl-r history search landing on a slash command no longer cancels the search
535- SDK: Support custom timeouts for hooks
536- Allow more safe git commands to run without approval
537- Plugins: Added support for sharing and installing output styles
538- Teleporting a session from web will automatically set the upstream branch
539
540## 2.0.37
541
542- Fixed how idleness is computed for notifications
543- Hooks: Added matcher values for Notification hook events
544- Output Styles: Added `keep-coding-instructions` option to frontmatter
545
546## 2.0.36
547
548- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
549- Fixed queued messages being incorrectly executed as bash commands
550- Fixed input being lost when typing while a queued message is processed
551
552## 2.0.35
553
554- Improve fuzzy search results when searching commands
555- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
556- 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
557- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
558- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
559
560## 2.0.34
561
562- VSCode Extension: Added setting to configure the initial permission mode for new conversations
563- Improved file path suggestion performance with native Rust-based fuzzy finder
564- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
565- Fixed memory crash when reading or writing large files (especially base64-encoded images)
566
567## 2.0.33
568
569- Native binary installs now launch quicker.
570- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
571- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
572
573## 2.0.32
574
575- Un-deprecate output styles based on community feedback
576- Added `companyAnnouncements` setting for displaying announcements on startup
577- Fixed hook progress messages not updating correctly during PostToolUse hook execution
578
579## 2.0.31
580
581- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
582- Vertex: add support for Web Search on supported models
583- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
584- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
585- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
586- Fixed plugin uninstall not removing plugins
587
588## 2.0.30
589
590- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
591- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
592- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
593- Added prompt-based stop hooks
594- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
595- Enabled SSE MCP servers on native build
596- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
597- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
598- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
599- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
600- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
601- Fixed bug that saved session permissions to local settings
602- Fixed MCP tools not being available to sub-agents
603- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
604- Fixed delay when navigating through typeahead suggestions with arrow keys
605- VSCode: Restored selection indicator in input footer showing current file or code selection status
606
607## 2.0.28
608
609- Plan mode: introduced new Plan subagent
610- Subagents: claude can now choose to resume subagents
611- Subagents: claude can dynamically choose the model used by its subagents
612- SDK: added --max-budget-usd flag
613- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
614- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
615- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
616- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
617- Various other bug fixes
618
619## 2.0.27
620
621- New UI for permission prompts
622- Added current branch filtering and search to session resume screen for easier navigation
623- Fixed directory @-mention causing "No assistant message found" error
624- VSCode Extension: Add config setting to include .gitignored files in file searches
625- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
626
627## 2.0.25
628
629- 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
630
631## 2.0.24
632
633- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
634- Claude Code Web: Support for Web -> CLI teleport
635- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
636- Bedrock: Display awsAuthRefresh output when auth is required
637
638## 2.0.22
639
640- Fixed content layout shift when scrolling through slash commands
641- IDE: Add toggle to enable/disable thinking.
642- Fix bug causing duplicate permission prompts with parallel tool calls
643- Add support for enterprise managed MCP allowlist and denylist
644
645## 2.0.21
646
647- Support MCP `structuredContent` field in tool responses
648- Added an interactive question tool
649- Claude will now ask you questions more often in plan mode
650- Added Haiku 4.5 as a model option for Pro users
651- Fixed an issue where queued commands don't have access to previous messages' output
652
653## 2.0.20
654
655- Added support for Claude Skills
656
657## 2.0.19
658
659- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
660- Fixed a bug where Haiku was unnecessarily called in print mode
661
662## 2.0.17
663
664- Added Haiku 4.5 to model selector!
665- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
666- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
667- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
668- OTEL: support HTTP_PROXY and HTTPS_PROXY
669- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
670
671## 2.0.15
672
673- Fixed bug with resuming where previously created files needed to be read again before writing
674- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
675
676## 2.0.14
677
678- Fix @-mentioning MCP servers to toggle them on/off
679- Improve permission checks for bash with inline env vars
680- Fix ultrathink + thinking toggle
681- Reduce unnecessary logins
682- Document --system-prompt
683- Several improvements to rendering
684- Plugins UI polish
685
686## 2.0.13
687
688- Fixed `/plugin` not working on native build
689
690## 2.0.12
691
692- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
693- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
694- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
695- `/plugin validate` command for validating plugin structure and configuration
696- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
697- Plugin documentation available at https://code.claude.com/docs/en/plugins
698- Comprehensive error messages and diagnostics via `/doctor` command
699- Avoid flickering in `/model` selector
700- Improvements to `/help`
701- Avoid mentioning hooks in `/resume` summaries
702- Changes to the "verbose" setting in `/config` now persist across sessions
703
704## 2.0.11
705
706- Reduced system prompt size by 1.4k tokens
707- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
708- Fixed Opus fallback rate limit errors appearing incorrectly
709- Fixed /add-dir command selecting wrong default tab
710
711## 2.0.10
712
713- Rewrote terminal renderer for buttery smooth UI
714- Enable/disable MCP servers by @mentioning, or in /mcp
715- Added tab completion for shell commands in bash mode
716- PreToolUse hooks can now modify tool inputs
717- Press Ctrl-G to edit your prompt in your system's configured text editor
718- Fixes for bash permission checks with environment variables in the command
719
720## 2.0.9
721
722- Fix regression where bash backgrounding stopped working
723
724## 2.0.8
725
726- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
727- IDE: Add drag-and-drop support for files and folders in chat
728- /context: Fix counting for thinking blocks
729- Improve message rendering for users with light themes on dark terminals
730- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
731
732## 2.0.5
733
734- IDE: Fix IME unintended message submission with Enter and Tab
735- IDE: Add "Open in Terminal" link in login screen
736- Fix unhandled OAuth expiration 401 API errors
737- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
738
739## 2.0.1
740
741- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
742- Various bug fixes and presentation improvements
743
744## 2.0.0
745
746- New native VS Code extension
747- Fresh coat of paint throughout the whole app
748- /rewind a conversation to undo code changes
749- /usage command to see plan limits
750- Tab to toggle thinking (sticky across sessions)
751- Ctrl-R to search history
752- Unshipped claude config command
753- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
754- SDK: The Claude Code SDK is now the Claude Agent SDK
755- Add subagents dynamically with `--agents` flag
756
757## 1.0.126
758
759- Enable /context command for Bedrock and Vertex
760- Add mTLS support for HTTP-based OpenTelemetry exporters
761
762## 1.0.124
763
764- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
765- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
766- No longer inform Claude of the list of allowed tools when permission is denied
767- Fixed security vulnerability in Bash tool permission checks
768- Improved VSCode extension performance for large files
769
770## 1.0.123
771
772- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
773- Fixed thinking mode triggering on negation phrases like "don't think"
774- Fixed rendering performance degradation during token streaming
775- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
776- Enhanced BashTool environment snapshot logging
777- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
778- Migrated --debug logging to a file, to enable easy tailing & filtering
779
780## 1.0.120
781
782- Fix input lag during typing, especially noticeable with large prompts
783- Improved VSCode extension command registry and sessions dialog user experience
784- Enhanced sessions dialog responsiveness and visual feedback
785- Fixed IDE compatibility issue by removing worktree support check
786- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
787
788## 1.0.119
789
790- Fix Windows issue where process visually freezes on entering interactive mode
791- Support dynamic headers for MCP servers via headersHelper configuration
792- Fix thinking mode not working in headless sessions
793- Fix slash commands now properly update allowed tools instead of replacing them
794
795## 1.0.117
796
797- Add Ctrl-R history search to recall previous commands like bash/zsh
798- Fix input lag while typing, especially on Windows
799- Add sed command to auto-allowed commands in acceptEdits mode
800- Fix Windows PATH comparison to be case-insensitive for drive letters
801- Add permissions management hint to /add-dir output
802
803## 1.0.115
804
805- Improve thinking mode display with enhanced visual effects
806- Type /t to temporarily disable thinking mode in your prompt
807- Improve path validation for glob and grep tools
808- Show condensed output for post-tool hooks to reduce visual clutter
809- Fix visual feedback when loading state completes
810- Improve UI consistency for permission request dialogs
811
812## 1.0.113
813
814- Deprecated piped input in interactive mode
815- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
816
817## 1.0.112
818
819- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
820- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
821- Hooks: Added systemMessage support for SessionEnd hooks
822- Added `spinnerTipsEnabled` setting to disable spinner tips
823- IDE: Various improvements and bug fixes
824
825## 1.0.111
826
827- /model now validates provided model names
828- Fixed Bash tool crashes caused by malformed shell syntax parsing
829
830## 1.0.110
831
832- /terminal-setup command now supports WezTerm
833- MCP: OAuth tokens now proactively refresh before expiration
834- Fixed reliability issues with background Bash processes
835
836## 1.0.109
837
838- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
839
840## 1.0.106
841
842- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
843
844## 1.0.97
845
846- Settings: /doctor now validates permission rule syntax and suggests corrections
847
848## 1.0.94
849
850- Vertex: add support for global endpoints for supported models
851- /memory command now allows direct editing of all imported memory files
852- SDK: Add custom tools as callbacks
853- Added /todos command to list current todo items
854
855## 1.0.93
856
857- Windows: Add alt + v shortcut for pasting images from clipboard
858- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
859
860## 1.0.90
861
862- Settings file changes take effect immediately - no restart required
863
864## 1.0.88
865
866- Fixed issue causing "OAuth authentication is currently not supported"
867- Status line input now includes `exceeds_200k_tokens`
868- Fixed incorrect usage tracking in /cost.
869- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
870- Bedrock: Updated default Sonnet model to Sonnet 4
871
872## 1.0.86
873
874- Added /context to help users self-serve debug context issues
875- SDK: Added UUID support for all SDK messages
876- SDK: Added `--replay-user-messages` to replay user messages back to stdout
877
878## 1.0.85
879
880- Status line input now includes session cost info
881- Hooks: Introduced SessionEnd hook
882
883## 1.0.84
884
885- Fix tool_use/tool_result id mismatch error when network is unstable
886- Fix Claude sometimes ignoring real-time steering when wrapping up a task
887- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
888- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
889
890## 1.0.83
891
892- @-mention: Support files with spaces in path
893- New shimmering spinner
894
895## 1.0.82
896
897- SDK: Add request cancellation support
898- SDK: New additionalDirectories option to search custom paths, improved slash command processing
899- Settings: Validation prevents invalid fields in .claude/settings.json files
900- MCP: Improve tool name consistency
901- Bash: Fix crash when Claude tries to automatically read large files
902
903## 1.0.81
904
905- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
906- Agents: Fix custom agent loading when agent files are unparsable
907
908## 1.0.80
909
910- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
911
912## 1.0.77
913
914- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
915- SDK: Add session support and permission denial tracking
916- Fix token limit errors in conversation summarization
917- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
918
919## 1.0.73
920
921- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
922- MCP: Press Esc to cancel OAuth authentication flows
923- Bash: Improved command validation and reduced false security warnings
924- UI: Enhanced spinner animations and status line visual hierarchy
925- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
926
927## 1.0.72
928
929- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
930
931## 1.0.71
932
933- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
934- Customizable status line: add your terminal prompt to Claude Code with /statusline
935
936## 1.0.70
937
938- Performance: Optimized message rendering for better performance with large contexts
939- Windows: Fixed native file search, ripgrep, and subagent functionality
940- Added support for @-mentions in slash command arguments
941
942## 1.0.69
943
944- Upgraded Opus to version 4.1
945
946## 1.0.68
947
948- Fix incorrect model names being used for certain commands like `/pr-comments`
949- 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.
950- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
951- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
952- SDK: Added canUseTool callback support for tool confirmation
953- Added `disableAllHooks` setting
954- Improved file suggestions performance in large repos
955
956## 1.0.65
957
958- IDE: Fixed connection stability issues and error handling for diagnostics
959- Windows: Fixed shell environment setup for users without .bashrc files
960
961## 1.0.64
962
963- Agents: Added model customization support - you can now specify which model an agent should use
964- Agents: Fixed unintended access to the recursive agent tool
965- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
966- SDK: Fixed user input tracking across multi-turn conversations
967- Added hidden files to file search and @-mention suggestions
968
969## 1.0.63
970
971- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
972
973## 1.0.62
974
975- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
976- Hooks: Added SessionStart hook for new session initialization
977- /add-dir command now supports typeahead for directory paths
978- Improved network connectivity check reliability
979
980## 1.0.61
981
982- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
983- Settings: Added `--settings` flag to load settings from a JSON file
984- Settings: Fixed resolution of settings files paths that are symlinks
985- OTEL: Fixed reporting of wrong organization after authentication changes
986- Slash commands: Fixed permissions checking for allowed-tools with Bash
987- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
988- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
989- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
990
991## 1.0.60
992
993- You can now create custom subagents for specialized tasks! Run /agents to get started
994
995## 1.0.59
996
997- SDK: Added tool confirmation support with canUseTool callback
998- SDK: Allow specifying env for spawned process
999- Hooks: Exposed PermissionDecision to hooks (including "ask")
1000- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1001- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1002
1003## 1.0.58
1004
1005- Added support for reading PDFs
1006- MCP: Improved server health status display in 'claude mcp list'
1007- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1008
1009## 1.0.57
1010
1011- Added support for specifying a model in slash commands
1012- Improved permission messages to help Claude understand allowed tools
1013- Fix: Remove trailing newlines from bash output in terminal wrapping
1014
1015## 1.0.56
1016
1017- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1018- Fixes for WSL IDE detection
1019- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1020
1021## 1.0.55
1022
1023- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1024- Windows: fixed Ctrl+Z crash
1025- SDK: Added ability to capture error logging
1026- Add --system-prompt-file option to override system prompt in print mode
1027
1028## 1.0.54
1029
1030- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1031- Custom slash commands: Added argument-hint to frontmatter
1032- Windows: OAuth uses port 45454 and properly constructs browser URL
1033- Windows: mode switching now uses alt + m, and plan mode renders properly
1034- Shell: Switch to in-memory shell snapshot to fix file-related errors
1035
1036## 1.0.53
1037
1038- Updated @-mention file truncation from 100 lines to 2000 lines
1039- 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).
1040
1041## 1.0.52
1042
1043- Added support for MCP server instructions
1044
1045## 1.0.51
1046
1047- Added support for native Windows (requires Git for Windows)
1048- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1049- Settings: /doctor can now help you identify and fix invalid setting files
1050- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1051- Increased auto-compact warning threshold from 60% to 80%
1052- Fixed an issue with handling user directories with spaces for shell snapshots
1053- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1054- Custom slash commands: Fixed user-level commands in subdirectories
1055- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1056
1057## 1.0.48
1058
1059- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1060- Added progress messages to Bash tool based on the last 5 lines of command output
1061- Added expanding variables support for MCP server configuration
1062- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1063- Improved IDE extension path handling when Claude Code runs in WSL
1064- Hooks: Added a PreCompact hook
1065- Vim mode: Added c, f/F, t/T
1066
1067## 1.0.45
1068
1069- Redesigned Search (Grep) tool with new tool input parameters and features
1070- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1071- Fixed config file corruption issue by enforcing atomic writes
1072- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1073- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1074- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1075
1076## 1.0.44
1077
1078- New /export command lets you quickly export a conversation for sharing
1079- MCP: resource_link tool results are now supported
1080- MCP: tool annotations and tool titles now display in /mcp view
1081- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1082
1083## 1.0.43
1084
1085- Fixed a bug where the theme selector was saving excessively
1086- Hooks: Added EPIPE system error handling
1087
1088## 1.0.42
1089
1090- Added tilde (`~`) expansion support to `/add-dir` command
1091
1092## 1.0.41
1093
1094- Hooks: Split Stop hook triggering into Stop and SubagentStop
1095- Hooks: Enabled optional timeout configuration for each command
1096- Hooks: Added "hook_event_name" to hook input
1097- Fixed a bug where MCP tools would display twice in tool list
1098- New tool parameters JSON for Bash tool in `tool_decision` event
1099
1100## 1.0.40
1101
1102- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1103
1104## 1.0.39
1105
1106- New Active Time metric in OpenTelemetry logging
1107
1108## 1.0.38
1109
1110- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1111
1112## 1.0.37
1113
1114- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1115
1116## 1.0.36
1117
1118- Web search now takes today's date into context
1119- Fixed a bug where stdio MCP servers were not terminating properly on exit
1120
1121## 1.0.35
1122
1123- Added support for MCP OAuth Authorization Server discovery
1124
1125## 1.0.34
1126
1127- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1128
1129## 1.0.33
1130
1131- Improved logging functionality with session ID support
1132- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1133- Improvements to plan mode
1134
1135## 1.0.32
1136
1137- Updated loopback config for litellm
1138- Added forceLoginMethod setting to bypass login selection screen
1139
1140## 1.0.31
1141
1142- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1143
1144## 1.0.30
1145
1146- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1147- Improved file path autocomplete with filename matching
1148- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1149- Enhanced jq regex support for complex filters with pipes and select
1150
1151## 1.0.29
1152
1153- Improved CJK character support in cursor navigation and rendering
1154
1155## 1.0.28
1156
1157- Slash commands: Fix selector display during history navigation
1158- Resizes images before upload to prevent API size limit errors
1159- Added XDG_CONFIG_HOME support to configuration directory
1160- Performance optimizations for memory usage
1161- New attributes (terminal.type, language) in OpenTelemetry logging
1162
1163## 1.0.27
1164
1165- Streamable HTTP MCP servers are now supported
1166- Remote MCP servers (SSE and HTTP) now support OAuth
1167- MCP resources can now be @-mentioned
1168- /resume slash command to switch conversations within Claude Code
1169
1170## 1.0.25
1171
1172- Slash commands: moved "project" and "user" prefixes to descriptions
1173- Slash commands: improved reliability for command discovery
1174- Improved support for Ghostty
1175- Improved web search reliability
1176
1177## 1.0.24
1178
1179- Improved /mcp output
1180- Fixed a bug where settings arrays got overwritten instead of merged
1181
1182## 1.0.23
1183
1184- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1185- Released Python SDK: pip install claude-code-sdk to get started
1186
1187## 1.0.22
1188
1189- SDK: Renamed `total_cost` to `total_cost_usd`
1190
1191## 1.0.21
1192
1193- Improved editing of files with tab-based indentation
1194- Fix for tool_use without matching tool_result errors
1195- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1196
1197## 1.0.18
1198
1199- Added --add-dir CLI argument for specifying additional working directories
1200- Added streaming input support without require -p flag
1201- Improved startup performance and session storage performance
1202- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1203- Added detailed MCP server tools display (/mcp)
1204- MCP authentication and permission improvements
1205- Added auto-reconnection for MCP SSE connections on disconnect
1206- Fixed issue where pasted content was lost when dialogs appeared
1207
1208## 1.0.17
1209
1210- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1211- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1212- MCP server list UI improvements
1213- Update Claude Code process title to display "claude" instead of "node"
1214
1215## 1.0.11
1216
1217- Claude Code can now also be used with a Claude Pro subscription
1218- Added /upgrade for smoother switching to Claude Max plans
1219- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1220- Improved shell configuration error handling
1221- Improved todo list handling during compaction
1222
1223## 1.0.10
1224
1225- Added markdown table support
1226- Improved streaming performance
1227
1228## 1.0.8
1229
1230- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1231- Increased default otel interval from 1s -> 5s
1232- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1233- Fixed a regression where search tools unnecessarily asked for permissions
1234- Added support for triggering thinking non-English languages
1235- Improved compacting UI
1236
1237## 1.0.7
1238
1239- Renamed /allowed-tools -> /permissions
1240- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1241- Deprecated claude config commands in favor of editing settings.json
1242- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1243- Improved error handling for /install-github-app
1244- Bugfixes, UI polish, and tool reliability improvements
1245
1246## 1.0.6
1247
1248- Improved edit reliability for tab-indented files
1249- Respect CLAUDE_CONFIG_DIR everywhere
1250- Reduced unnecessary tool permission prompts
1251- Added support for symlinks in @file typeahead
1252- Bugfixes, UI polish, and tool reliability improvements
1253
1254## 1.0.4
1255
1256- Fixed a bug where MCP tool errors weren't being parsed correctly
1257
1258## 1.0.1
1259
1260- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1261- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1262- Updated documentation links and OAuth process descriptions
1263
1264## 1.0.0
1265
1266- Claude Code is now generally available
1267- Introducing Sonnet 4 and Opus 4 models
1268
1269## 0.2.125
1270
1271- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1272- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1273
1274## 0.2.117
1275
1276- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1277- Introduced settings.cleanupPeriodDays
1278- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1279- Introduced --debug mode
1280
1281## 0.2.108
1282
1283- You can now send messages to Claude while it works to steer Claude in real-time
1284- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1285- Fixed a bug where thinking was not working in -p mode
1286- Fixed a regression in /cost reporting
1287- Deprecated MCP wizard interface in favor of other MCP commands
1288- Lots of other bugfixes and improvements
1289
1290## 0.2.107
1291
1292- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1293
1294## 0.2.106
1295
1296- MCP SSE server configs can now specify custom headers
1297- Fixed a bug where MCP permission prompt didn't always show correctly
1298
1299## 0.2.105
1300
1301- Claude can now search the web
1302- Moved system & account status to /status
1303- Added word movement keybindings for Vim
1304- Improved latency for startup, todo tool, and file edits
1305
1306## 0.2.102
1307
1308- Improved thinking triggering reliability
1309- Improved @mention reliability for images and folders
1310- You can now paste multiple large chunks into one prompt
1311
1312## 0.2.100
1313
1314- Fixed a crash caused by a stack overflow error
1315- Made db storage optional; missing db support disables --continue and --resume
1316
1317## 0.2.98
1318
1319- Fixed an issue where auto-compact was running twice
1320
1321## 0.2.96
1322
1323- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1324
1325## 0.2.93
1326
1327- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1328- Claude now has access to a Todo list that helps it stay on track and be more organized
1329
1330## 0.2.82
1331
1332- Added support for --disallowedTools
1333- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1334
1335## 0.2.75
1336
1337- Hit Enter to queue up additional messages while Claude is working
1338- Drag in or copy/paste image files directly into the prompt
1339- @-mention files to directly add them to context
1340- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1341- Improved performance for filename auto-complete
1342
1343## 0.2.74
1344
1345- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1346- Task tool can now perform writes and run bash commands
1347
1348## 0.2.72
1349
1350- Updated spinner to indicate tokens loaded and tool usage
1351
1352## 0.2.70
1353
1354- Network commands like curl are now available for Claude to use
1355- Claude can now run multiple web queries in parallel
1356- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1357
1358## 0.2.69
1359
1360- Fixed UI glitches with improved Select component behavior
1361- Enhanced terminal output display with better text truncation logic
1362
1363## 0.2.67
1364
1365- Shared project permission rules can be saved in .claude/settings.json
1366
1367## 0.2.66
1368
1369- Print mode (-p) now supports streaming output via --output-format=stream-json
1370- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1371
1372## 0.2.63
1373
1374- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1375
1376## 0.2.61
1377
1378- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1379- Enhanced image detection for more reliable clipboard paste functionality
1380- Fixed an issue where ESC key could crash the conversation history selector
1381
1382## 0.2.59
1383
1384- Copy+paste images directly into your prompt
1385- Improved progress indicators for bash and fetch tools
1386- Bugfixes for non-interactive mode (-p)
1387
1388## 0.2.54
1389
1390- Quickly add to Memory by starting your message with '#'
1391- Press ctrl+r to see full output for long tool results
1392- Added support for MCP SSE transport
1393
1394## 0.2.53
1395
1396- New web fetch tool lets Claude view URLs that you paste in
1397- Fixed a bug with JPEG detection
1398
1399## 0.2.50
1400
1401- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1402
1403## 0.2.49
1404
1405- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1406
1407## 0.2.47
1408
1409- Press Tab to auto-complete file and folder names
1410- Press Shift + Tab to toggle auto-accept for file edits
1411- Automatic conversation compaction for infinite conversation length (toggle with /config)
1412
1413## 0.2.44
1414
1415- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1416
1417## 0.2.41
1418
1419- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1420- MCP server startup no longer blocks the app from starting up
1421
1422## 0.2.37
1423
1424- New /release-notes command lets you view release notes at any time
1425- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1426
1427## 0.2.36
1428
1429- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1430- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1431
1432## 0.2.34
1433
1434- Vim bindings for text input - enable with /vim or /config
1435
1436## 0.2.32
1437
1438- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1439- Fix for some PersistentShell issues
1440
1441## 0.2.31
1442
1443- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1444- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1445
1446## 0.2.30
1447
1448- Added ANSI color theme for better terminal compatibility
1449- Fixed issue where slash command arguments weren't being sent properly
1450- (Mac-only) API keys are now stored in macOS Keychain
1451
1452## 0.2.26
1453
1454- New /approved-tools command for managing tool permissions
1455- Word-level diff display for improved code readability
1456- Fuzzy matching for slash commands
1457
1458## 0.2.21
1459
1460- Fuzzy matching for /commands