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 8c48d2f · on Jan 28, 2026 History
1,529 lines · 78.0 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.21
4
5- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
6- Fixed shell completion cache files being truncated on exit
7- Fixed API errors when resuming sessions that were interrupted during tool execution
8- Fixed auto-compact triggering too early on models with large output token limits
9- Fixed task IDs potentially being reused after deletion
10- Fixed file search not working in VS Code extension on Windows
11- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
12- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
13- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
14- [VSCode] Fixed message action buttons having incorrect background colors
15
16## 2.1.20
17
18- Added arrow key history navigation in vim normal mode when cursor cannot move further
19- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
20- Added PR review status indicator to the prompt footer, showing the current branch's PR state (approved, changes requested, pending, or draft) as a colored dot with a clickable link
21- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
22- Added ability to delete tasks via the `TaskUpdate` tool
23- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
24- Fixed agents sometimes ignoring user messages sent while actively working on a task
25- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
26- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
27- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
28- Fixed draft prompt being lost when pressing UP arrow to navigate command history
29- Fixed ghost text flickering when typing slash commands mid-input
30- Fixed marketplace source removal not properly deleting settings
31- Fixed duplicate output in some commands like `/context`
32- Fixed task list sometimes showing outside the main conversation view
33- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
34- Fixed crashes when cancelling tool use
35- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
36- Improved thinking status text with a subtle shimmer animation
37- Improved task list to dynamically adjust visible items based on terminal height
38- Improved fork conversation hint to show how to resume the original session
39- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
40- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
41- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
42- Changed the `/copy` command to be available to all users
43- Changed background agents to prompt for tool permissions before launching
44- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
45- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
46
47## 2.1.19
48
49- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
50- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
51- Fixed crashes on processors without AVX instruction support
52- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
53- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
54- Fixed resuming sessions by custom title not working when run from a different directory
55- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
56- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
57- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
58- Fixed file write preview omitting empty lines
59- Changed skills without additional permissions or hooks to be allowed without requiring approval
60- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
61- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
62- [VSCode] Enabled session forking and rewind functionality for all users
63
64## 2.1.18
65
66- Added customizable keyboard shortcuts. Configure keybindings per context, create chord sequences, and personalize your workflow. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
67
68## 2.1.17
69
70- Fixed crashes on processors without AVX instruction support
71
72## 2.1.16
73
74- Added new task management system, including new capabilities like dependency tracking
75- [VSCode] Added native plugin management support
76- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
77- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
78- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
79- Fixed session titles on the resume screen not respecting the user's language setting
80- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
81
82## 2.1.15
83
84- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
85- Improved UI rendering performance with React Compiler
86- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
87- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
88
89## 2.1.14
90
91- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
92- Added search to installed plugins list - type to filter by name or description
93- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
94- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
95- Fixed memory issues that could cause crashes when running parallel subagents
96- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
97- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
98- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
99- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
100- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
101- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
102- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
103- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
104- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
105- Improved backspace to delete pasted text as a single token instead of one character at a time
106- [VSCode] Added `/usage` command to display current plan usage
107
108## 2.1.12
109
110- Fixed message rendering bug
111
112## 2.1.11
113
114- Fixed excessive MCP connection requests for HTTP/SSE transports
115
116## 2.1.10
117
118- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
119- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
120- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
121- Improved startup to capture keystrokes typed before the REPL is fully ready
122- Improved file suggestions to show as removable attachments instead of inserting text when accepted
123- [VSCode] Added install count display to plugin listings
124- [VSCode] Added trust warning when installing plugins
125
126## 2.1.9
127
128- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
129- Added `plansDirectory` setting to customize where plan files are stored
130- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
131- Added session URL attribution to commits and PRs created from web sessions
132- Added support for `PreToolUse` hooks to return `additionalContext` to the model
133- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
134- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
135- Fixed MCP server reconnection hanging when cached connection promise never resolves
136- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
137
138## 2.1.7
139
140- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
141- Added ability to provide feedback when accepting permission prompts
142- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
143- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
144- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
145- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
146- 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)
147- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
148- Fixed terminal title animation jitter by using fixed-width braille characters
149- Fixed plugins with git submodules not being fully initialized when installed
150- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
151- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
152- 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.
153- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
154- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
155
156## 2.1.6
157
158- Added search functionality to `/config` command for quickly filtering settings
159- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
160- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
161- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
162- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
163- Added an error display when the editor fails during Ctrl+G
164- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
165- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
166- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
167- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
168- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
169- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
170- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
171- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
172- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
173- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
174- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
175- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
176- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
177- Improved @ autocomplete with icons for different suggestion types and single-line formatting
178- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
179- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
180- Changed terminal title to "Claude Code" on startup for better window identification
181- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
182- [VSCode] Fixed usage indicator not updating after manual compact
183
184## 2.1.5
185
186- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
187
188## 2.1.4
189
190- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
191- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
192
193## 2.1.3
194
195- Merged slash commands and skills, simplifying the mental model with no change in behavior
196- Added release channel (`stable` or `latest`) toggle to `/config`
197- 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
198- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
199- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
200- Fixed mismatch between background task count in status bar and items shown in tasks dialog
201- Fixed sub-agents using the wrong model during conversation compaction
202- Fixed web search in sub-agents using incorrect model
203- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
204- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
205- Improved slash command suggestion readability by truncating long descriptions to 2 lines
206- Changed tool hook execution timeout from 60 seconds to 10 minutes
207- [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)
208
209## 2.1.2
210
211- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
212- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
213- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
214- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
215- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
216- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
217- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
218- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
219- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
220- Fixed updates incorrectly claiming another installation is in progress
221- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
222- Fixed remote session URL and teleport being broken when using `/tasks` command
223- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
224- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
225- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
226- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
227- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
228- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
229- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
230- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
231- [SDK] Changed minimum zod peer dependency to ^4.0.0
232- [VSCode] Fixed usage display not updating after manual compact
233
234## 2.1.0
235
236- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
237- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
238- Added support for `agent` field in skills to specify agent type for execution
239- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
240- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
241- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
242- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
243- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
244- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
245- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
246- Fixed Esc key with queued prompts to only move them to input without canceling the running task
247- Reduced permission prompts for complex bash commands
248- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
249- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
250- Fixed plugin path resolution for file-based marketplace sources
251- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
252- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
253- Fixed Claude in Chrome support for WSL environments
254- Fixed Windows native installer silently failing when executable creation fails
255- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
256- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
257- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
258- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
259- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
260- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
261- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
262- Added hooks support for skill and slash command frontmatter
263- 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
264- Added `/plan` command shortcut to enable plan mode directly from the prompt
265- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
266- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
267- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
268- Added support for `once: true` config for hooks
269- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
270- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
271- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
272- Added left/right arrow key navigation for cycling through tabs in dialogs
273- Added real-time thinking block display in Ctrl+O transcript mode
274- Added filepath to full output in background bash task details dialog
275- Added Skills as a separate category in the context visualization
276- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
277- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
278- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
279- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
280- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
281- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
282- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
283- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
284- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
285- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
286- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
287- Fixed commands with `$()` command substitution failing with parse errors
288- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
289- 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)
290- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
291- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
292- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
293- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
294- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
295- Fixed subagents sometimes not inheriting the parent's model by default
296- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
297- Fixed duplicate Bash commands appearing in permission request option labels
298- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
299- Fixed background task completion notifications to appear proactively with bullet point
300- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
301- Fixed cursor disappearing after dismissing permission dialogs
302- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
303- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
304- Fixed images being silently dropped when queueing messages while backgrounding a task
305- Fixed large pasted images failing with "Image was too large" error
306- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
307- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
308- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
309- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
310- Fixed spinner token counter to properly accumulate tokens from subagents during execution
311- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
312- Fixed race condition where LSP tool could return "no server available" during startup
313- Fixed feedback submission hanging indefinitely when network requests timeout
314- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
315- Fixed hook success message showing trailing colon when hook has no output
316- Multiple optimizations to improve startup performance
317- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
318- Improved performance when reading Jupyter notebooks with many cells
319- Improved reliability for piped input like `cat refactor.md | claude`
320- Improved reliability for AskQuestion tool
321- Improved sed in-place edit commands to render as file edits with diff preview
322- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
323- Improved compaction reliability
324- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
325- Improved skills to show progress while executing, displaying tool uses as they happen
326- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
327- Improved skill suggestions to prioritize recently and frequently used skills
328- Improved spinner feedback when waiting for the first response token
329- Improved token count display in spinner to include tokens from background agents
330- Improved incremental output for async agents to give the main thread more control and visibility
331- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
332- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
333- Improved macOS screenshot paste reliability with TIFF format support
334- Improved `/stats` output
335- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
336- Changed "Interrupted" message color from red to grey for a less alarming appearance
337- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
338- Removed underline styling from image reference links
339- [SDK] Changed minimum zod peer dependency to ^4.0.0
340- [VSCode] Added currently selected model name to the context menu
341- [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")
342- [VSCode] Fixed paragraph breaks not rendering in markdown content
343- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
344- [Windows] Fixed issue with improper rendering
345
346## 2.0.76
347
348- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
349
350## 2.0.75
351
352- Minor bugfixes
353
354## 2.0.74
355
356- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
357- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
358- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
359- Added syntax highlighting info to theme picker
360- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
361- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
362- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
363- Fixed a potential crash when syntax highlighting isn't initialized correctly
364- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
365- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
366- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
367- [Windows] Fixed issue with improper rendering
368- [VSCode] Added gift tag pictogram for year-end promotion message
369
370## 2.0.73
371
372- Added clickable `[Image #N]` links that open attached images in the default viewer
373- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
374- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
375- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
376- Fixed slow input history cycling and race condition that could overwrite text after message submission
377- Improved `/theme` command to open theme picker directly
378- Improved theme picker UI
379- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
380- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
381
382## 2.0.72
383
384- 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
385- Reduced terminal flickering
386- Added scannable QR code to mobile app tip for quick app downloads
387- Added loading indicator when resuming conversations for better feedback
388- Fixed `/context` command not respecting custom system prompts in non-interactive mode
389- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
390- Improved @ mention file suggestion speed (~3x faster in git repositories)
391- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
392- Improved settings validation errors to be more prominent
393- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
394
395## 2.0.71
396
397- Added /config toggle to enable/disable prompt suggestions
398- Added `/settings` as an alias for the `/config` command
399- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
400- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
401- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
402- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
403- New syntax highlighting engine for native build
404
405## 2.0.70
406
407- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
408- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
409- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
410- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
411- Fixed input being cleared when processing queued commands while the user was typing
412- Fixed prompt suggestions replacing typed input when pressing Tab
413- Fixed diff view not updating when terminal is resized
414- Improved memory usage by 3x for large conversations
415- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
416- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
417- Fix thinking mode toggle in /config not persisting correctly
418- Improve UI for file creation permission dialog
419
420## 2.0.69
421
422- Minor bugfixes
423
424## 2.0.68
425
426- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
427- Fixed a bug where disallowed MCP tools were visible to the model
428- Fixed an issue where steering messages could be lost while a subagent is working
429- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
430- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
431- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
432
433## 2.0.67
434
435- Thinking mode is now enabled by default for Opus 4.5
436- Thinking mode configuration has moved to /config
437- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
438- Show reason why autoupdater is disabled in `/doctor`
439- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
440- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
441- Fixed scroll position resetting after deleting a permission rule in `/permissions`
442- 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
443- Fixed `claude install --force` not bypassing stale lock files
444- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
445- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
446
447## 2.0.65
448
449- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
450- Added context window information to status line input
451- Added `fileSuggestion` setting for custom `@` file search commands
452- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
453- Fixed prompt not being saved to history when aborting a query with Escape
454- Fixed Read tool image handling to identify format from bytes instead of file extension
455
456## 2.0.64
457
458- Made auto-compacting instant
459- Agents and bash commands can run asynchronously and send messages to wake up the main agent
460- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
461- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
462- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
463- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
464- Fixed auto-loading .env when using native installer
465- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
466- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
467- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
468- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
469- Bedrock: Improve efficiency of token counting
470- Bedrock: Add support for `aws login` AWS Management Console credentials
471- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
472
473## 2.0.62
474
475- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
476- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
477- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
478- Fixed slash command selection not working when multiple commands share the same name
479- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
480- Fixed running versions getting removed because lock file incorrectly going stale
481- Fixed IDE diff tab not closing when rejecting file changes
482
483## 2.0.61
484
485- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
486
487## 2.0.60
488
489- Added background agent support. Agents run in the background while you work
490- Added --disable-slash-commands CLI flag to disable all slash commands
491- Added model name to "Co-Authored-By" commit messages
492- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
493- Updated Fetch to skip summarization for pre-approved websites
494- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
495
496## 2.0.59
497
498- Added --agent CLI flag to override the agent setting for the current session
499- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
500- VS Code: Fixed .claude.json config file being read from incorrect location
501
502## 2.0.58
503
504- Pro users now have access to Opus 4.5 as part of their subscription!
505- Fixed timer duration showing "11m 60s" instead of "12m 0s"
506- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
507
508## 2.0.57
509
510- Added feedback input when rejecting plans, allowing users to tell Claude what to change
511- VSCode: Added streaming message support for real-time response display
512
513## 2.0.56
514
515- Added setting to enable/disable terminal progress bar (OSC 9;4)
516- 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.
517
518## 2.0.55
519
520- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
521- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
522- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
523- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
524
525## 2.0.54
526
527- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
528- Fix issue with excessive iTerm notifications
529
530## 2.0.52
531
532- Fixed duplicate message display when starting Claude with a command line argument
533- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
534- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
535- Permit some uses of `$!` in bash commands
536
537## 2.0.51
538
539- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
540- Introducing Claude Code for Desktop: https://claude.com/download
541- 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
542- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
543- Plan Mode now builds more precise plans and executes more thoroughly
544- Usage limit notifications now easier to understand
545- Switched `/usage` back to "% used"
546- Fixed handling of thinking errors
547- Fixed performance regression
548
549## 2.0.50
550
551- Fixed bug preventing calling MCP tools that have nested references in their input schemas
552- Silenced a noisy but harmless error during upgrades
553- Improved ultrathink text display
554- Improved clarity of 5-hour session limit warning message
555
556## 2.0.49
557
558- Added readline-style ctrl-y for pasting deleted text
559- Improved clarity of usage limit warning message
560- Fixed handling of subagent permissions
561
562## 2.0.47
563
564- Improved error messages and validation for `claude --teleport`
565- Improved error handling in `/usage`
566- Fixed race condition with history entry not getting logged at exit
567- Fixed Vertex AI configuration not being applied from `settings.json`
568
569## 2.0.46
570
571- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
572
573## 2.0.45
574
575- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
576- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
577- Send background tasks to Claude Code on the web by starting a message with `&`
578
579## 2.0.43
580
581- Added `permissionMode` field for custom agents
582- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
583- Added skills frontmatter field to declare skills to auto-load for subagents
584- Added the `SubagentStart` hook event
585- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
586- Fixed duplicate rendering of some messages in the UI
587- Fixed some visual flickers
588- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
589
590## 2.0.42
591
592- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
593
594## 2.0.41
595
596- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
597- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
598- Fixed incorrect labeling of user settings vs project settings in command descriptions
599- Fixed crash when plugin command hooks timeout during execution
600- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
601- Fixed broken security documentation links in trust dialogs and onboarding
602- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
603- ctrl-r history search landing on a slash command no longer cancels the search
604- SDK: Support custom timeouts for hooks
605- Allow more safe git commands to run without approval
606- Plugins: Added support for sharing and installing output styles
607- Teleporting a session from web will automatically set the upstream branch
608
609## 2.0.37
610
611- Fixed how idleness is computed for notifications
612- Hooks: Added matcher values for Notification hook events
613- Output Styles: Added `keep-coding-instructions` option to frontmatter
614
615## 2.0.36
616
617- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
618- Fixed queued messages being incorrectly executed as bash commands
619- Fixed input being lost when typing while a queued message is processed
620
621## 2.0.35
622
623- Improve fuzzy search results when searching commands
624- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
625- 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
626- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
627- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
628
629## 2.0.34
630
631- VSCode Extension: Added setting to configure the initial permission mode for new conversations
632- Improved file path suggestion performance with native Rust-based fuzzy finder
633- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
634- Fixed memory crash when reading or writing large files (especially base64-encoded images)
635
636## 2.0.33
637
638- Native binary installs now launch quicker.
639- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
640- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
641
642## 2.0.32
643
644- Un-deprecate output styles based on community feedback
645- Added `companyAnnouncements` setting for displaying announcements on startup
646- Fixed hook progress messages not updating correctly during PostToolUse hook execution
647
648## 2.0.31
649
650- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
651- Vertex: add support for Web Search on supported models
652- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
653- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
654- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
655- Fixed plugin uninstall not removing plugins
656
657## 2.0.30
658
659- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
660- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
661- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
662- Added prompt-based stop hooks
663- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
664- Enabled SSE MCP servers on native build
665- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
666- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
667- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
668- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
669- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
670- Fixed bug that saved session permissions to local settings
671- Fixed MCP tools not being available to sub-agents
672- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
673- Fixed delay when navigating through typeahead suggestions with arrow keys
674- VSCode: Restored selection indicator in input footer showing current file or code selection status
675
676## 2.0.28
677
678- Plan mode: introduced new Plan subagent
679- Subagents: claude can now choose to resume subagents
680- Subagents: claude can dynamically choose the model used by its subagents
681- SDK: added --max-budget-usd flag
682- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
683- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
684- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
685- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
686- Various other bug fixes
687
688## 2.0.27
689
690- New UI for permission prompts
691- Added current branch filtering and search to session resume screen for easier navigation
692- Fixed directory @-mention causing "No assistant message found" error
693- VSCode Extension: Add config setting to include .gitignored files in file searches
694- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
695
696## 2.0.25
697
698- 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
699
700## 2.0.24
701
702- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
703- Claude Code Web: Support for Web -> CLI teleport
704- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
705- Bedrock: Display awsAuthRefresh output when auth is required
706
707## 2.0.22
708
709- Fixed content layout shift when scrolling through slash commands
710- IDE: Add toggle to enable/disable thinking.
711- Fix bug causing duplicate permission prompts with parallel tool calls
712- Add support for enterprise managed MCP allowlist and denylist
713
714## 2.0.21
715
716- Support MCP `structuredContent` field in tool responses
717- Added an interactive question tool
718- Claude will now ask you questions more often in plan mode
719- Added Haiku 4.5 as a model option for Pro users
720- Fixed an issue where queued commands don't have access to previous messages' output
721
722## 2.0.20
723
724- Added support for Claude Skills
725
726## 2.0.19
727
728- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
729- Fixed a bug where Haiku was unnecessarily called in print mode
730
731## 2.0.17
732
733- Added Haiku 4.5 to model selector!
734- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
735- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
736- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
737- OTEL: support HTTP_PROXY and HTTPS_PROXY
738- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
739
740## 2.0.15
741
742- Fixed bug with resuming where previously created files needed to be read again before writing
743- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
744
745## 2.0.14
746
747- Fix @-mentioning MCP servers to toggle them on/off
748- Improve permission checks for bash with inline env vars
749- Fix ultrathink + thinking toggle
750- Reduce unnecessary logins
751- Document --system-prompt
752- Several improvements to rendering
753- Plugins UI polish
754
755## 2.0.13
756
757- Fixed `/plugin` not working on native build
758
759## 2.0.12
760
761- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
762- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
763- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
764- `/plugin validate` command for validating plugin structure and configuration
765- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
766- Plugin documentation available at https://code.claude.com/docs/en/plugins
767- Comprehensive error messages and diagnostics via `/doctor` command
768- Avoid flickering in `/model` selector
769- Improvements to `/help`
770- Avoid mentioning hooks in `/resume` summaries
771- Changes to the "verbose" setting in `/config` now persist across sessions
772
773## 2.0.11
774
775- Reduced system prompt size by 1.4k tokens
776- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
777- Fixed Opus fallback rate limit errors appearing incorrectly
778- Fixed /add-dir command selecting wrong default tab
779
780## 2.0.10
781
782- Rewrote terminal renderer for buttery smooth UI
783- Enable/disable MCP servers by @mentioning, or in /mcp
784- Added tab completion for shell commands in bash mode
785- PreToolUse hooks can now modify tool inputs
786- Press Ctrl-G to edit your prompt in your system's configured text editor
787- Fixes for bash permission checks with environment variables in the command
788
789## 2.0.9
790
791- Fix regression where bash backgrounding stopped working
792
793## 2.0.8
794
795- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
796- IDE: Add drag-and-drop support for files and folders in chat
797- /context: Fix counting for thinking blocks
798- Improve message rendering for users with light themes on dark terminals
799- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
800
801## 2.0.5
802
803- IDE: Fix IME unintended message submission with Enter and Tab
804- IDE: Add "Open in Terminal" link in login screen
805- Fix unhandled OAuth expiration 401 API errors
806- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
807
808## 2.0.1
809
810- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
811- Various bug fixes and presentation improvements
812
813## 2.0.0
814
815- New native VS Code extension
816- Fresh coat of paint throughout the whole app
817- /rewind a conversation to undo code changes
818- /usage command to see plan limits
819- Tab to toggle thinking (sticky across sessions)
820- Ctrl-R to search history
821- Unshipped claude config command
822- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
823- SDK: The Claude Code SDK is now the Claude Agent SDK
824- Add subagents dynamically with `--agents` flag
825
826## 1.0.126
827
828- Enable /context command for Bedrock and Vertex
829- Add mTLS support for HTTP-based OpenTelemetry exporters
830
831## 1.0.124
832
833- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
834- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
835- No longer inform Claude of the list of allowed tools when permission is denied
836- Fixed security vulnerability in Bash tool permission checks
837- Improved VSCode extension performance for large files
838
839## 1.0.123
840
841- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
842- Fixed thinking mode triggering on negation phrases like "don't think"
843- Fixed rendering performance degradation during token streaming
844- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
845- Enhanced BashTool environment snapshot logging
846- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
847- Migrated --debug logging to a file, to enable easy tailing & filtering
848
849## 1.0.120
850
851- Fix input lag during typing, especially noticeable with large prompts
852- Improved VSCode extension command registry and sessions dialog user experience
853- Enhanced sessions dialog responsiveness and visual feedback
854- Fixed IDE compatibility issue by removing worktree support check
855- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
856
857## 1.0.119
858
859- Fix Windows issue where process visually freezes on entering interactive mode
860- Support dynamic headers for MCP servers via headersHelper configuration
861- Fix thinking mode not working in headless sessions
862- Fix slash commands now properly update allowed tools instead of replacing them
863
864## 1.0.117
865
866- Add Ctrl-R history search to recall previous commands like bash/zsh
867- Fix input lag while typing, especially on Windows
868- Add sed command to auto-allowed commands in acceptEdits mode
869- Fix Windows PATH comparison to be case-insensitive for drive letters
870- Add permissions management hint to /add-dir output
871
872## 1.0.115
873
874- Improve thinking mode display with enhanced visual effects
875- Type /t to temporarily disable thinking mode in your prompt
876- Improve path validation for glob and grep tools
877- Show condensed output for post-tool hooks to reduce visual clutter
878- Fix visual feedback when loading state completes
879- Improve UI consistency for permission request dialogs
880
881## 1.0.113
882
883- Deprecated piped input in interactive mode
884- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
885
886## 1.0.112
887
888- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
889- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
890- Hooks: Added systemMessage support for SessionEnd hooks
891- Added `spinnerTipsEnabled` setting to disable spinner tips
892- IDE: Various improvements and bug fixes
893
894## 1.0.111
895
896- /model now validates provided model names
897- Fixed Bash tool crashes caused by malformed shell syntax parsing
898
899## 1.0.110
900
901- /terminal-setup command now supports WezTerm
902- MCP: OAuth tokens now proactively refresh before expiration
903- Fixed reliability issues with background Bash processes
904
905## 1.0.109
906
907- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
908
909## 1.0.106
910
911- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
912
913## 1.0.97
914
915- Settings: /doctor now validates permission rule syntax and suggests corrections
916
917## 1.0.94
918
919- Vertex: add support for global endpoints for supported models
920- /memory command now allows direct editing of all imported memory files
921- SDK: Add custom tools as callbacks
922- Added /todos command to list current todo items
923
924## 1.0.93
925
926- Windows: Add alt + v shortcut for pasting images from clipboard
927- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
928
929## 1.0.90
930
931- Settings file changes take effect immediately - no restart required
932
933## 1.0.88
934
935- Fixed issue causing "OAuth authentication is currently not supported"
936- Status line input now includes `exceeds_200k_tokens`
937- Fixed incorrect usage tracking in /cost.
938- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
939- Bedrock: Updated default Sonnet model to Sonnet 4
940
941## 1.0.86
942
943- Added /context to help users self-serve debug context issues
944- SDK: Added UUID support for all SDK messages
945- SDK: Added `--replay-user-messages` to replay user messages back to stdout
946
947## 1.0.85
948
949- Status line input now includes session cost info
950- Hooks: Introduced SessionEnd hook
951
952## 1.0.84
953
954- Fix tool_use/tool_result id mismatch error when network is unstable
955- Fix Claude sometimes ignoring real-time steering when wrapping up a task
956- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
957- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
958
959## 1.0.83
960
961- @-mention: Support files with spaces in path
962- New shimmering spinner
963
964## 1.0.82
965
966- SDK: Add request cancellation support
967- SDK: New additionalDirectories option to search custom paths, improved slash command processing
968- Settings: Validation prevents invalid fields in .claude/settings.json files
969- MCP: Improve tool name consistency
970- Bash: Fix crash when Claude tries to automatically read large files
971
972## 1.0.81
973
974- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
975- Agents: Fix custom agent loading when agent files are unparsable
976
977## 1.0.80
978
979- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
980
981## 1.0.77
982
983- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
984- SDK: Add session support and permission denial tracking
985- Fix token limit errors in conversation summarization
986- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
987
988## 1.0.73
989
990- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
991- MCP: Press Esc to cancel OAuth authentication flows
992- Bash: Improved command validation and reduced false security warnings
993- UI: Enhanced spinner animations and status line visual hierarchy
994- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
995
996## 1.0.72
997
998- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
999
1000## 1.0.71
1001
1002- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1003- Customizable status line: add your terminal prompt to Claude Code with /statusline
1004
1005## 1.0.70
1006
1007- Performance: Optimized message rendering for better performance with large contexts
1008- Windows: Fixed native file search, ripgrep, and subagent functionality
1009- Added support for @-mentions in slash command arguments
1010
1011## 1.0.69
1012
1013- Upgraded Opus to version 4.1
1014
1015## 1.0.68
1016
1017- Fix incorrect model names being used for certain commands like `/pr-comments`
1018- 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.
1019- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1020- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1021- SDK: Added canUseTool callback support for tool confirmation
1022- Added `disableAllHooks` setting
1023- Improved file suggestions performance in large repos
1024
1025## 1.0.65
1026
1027- IDE: Fixed connection stability issues and error handling for diagnostics
1028- Windows: Fixed shell environment setup for users without .bashrc files
1029
1030## 1.0.64
1031
1032- Agents: Added model customization support - you can now specify which model an agent should use
1033- Agents: Fixed unintended access to the recursive agent tool
1034- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1035- SDK: Fixed user input tracking across multi-turn conversations
1036- Added hidden files to file search and @-mention suggestions
1037
1038## 1.0.63
1039
1040- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1041
1042## 1.0.62
1043
1044- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1045- Hooks: Added SessionStart hook for new session initialization
1046- /add-dir command now supports typeahead for directory paths
1047- Improved network connectivity check reliability
1048
1049## 1.0.61
1050
1051- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1052- Settings: Added `--settings` flag to load settings from a JSON file
1053- Settings: Fixed resolution of settings files paths that are symlinks
1054- OTEL: Fixed reporting of wrong organization after authentication changes
1055- Slash commands: Fixed permissions checking for allowed-tools with Bash
1056- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1057- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1058- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1059
1060## 1.0.60
1061
1062- You can now create custom subagents for specialized tasks! Run /agents to get started
1063
1064## 1.0.59
1065
1066- SDK: Added tool confirmation support with canUseTool callback
1067- SDK: Allow specifying env for spawned process
1068- Hooks: Exposed PermissionDecision to hooks (including "ask")
1069- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1070- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1071
1072## 1.0.58
1073
1074- Added support for reading PDFs
1075- MCP: Improved server health status display in 'claude mcp list'
1076- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1077
1078## 1.0.57
1079
1080- Added support for specifying a model in slash commands
1081- Improved permission messages to help Claude understand allowed tools
1082- Fix: Remove trailing newlines from bash output in terminal wrapping
1083
1084## 1.0.56
1085
1086- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1087- Fixes for WSL IDE detection
1088- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1089
1090## 1.0.55
1091
1092- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1093- Windows: fixed Ctrl+Z crash
1094- SDK: Added ability to capture error logging
1095- Add --system-prompt-file option to override system prompt in print mode
1096
1097## 1.0.54
1098
1099- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1100- Custom slash commands: Added argument-hint to frontmatter
1101- Windows: OAuth uses port 45454 and properly constructs browser URL
1102- Windows: mode switching now uses alt + m, and plan mode renders properly
1103- Shell: Switch to in-memory shell snapshot to fix file-related errors
1104
1105## 1.0.53
1106
1107- Updated @-mention file truncation from 100 lines to 2000 lines
1108- 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).
1109
1110## 1.0.52
1111
1112- Added support for MCP server instructions
1113
1114## 1.0.51
1115
1116- Added support for native Windows (requires Git for Windows)
1117- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1118- Settings: /doctor can now help you identify and fix invalid setting files
1119- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1120- Increased auto-compact warning threshold from 60% to 80%
1121- Fixed an issue with handling user directories with spaces for shell snapshots
1122- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1123- Custom slash commands: Fixed user-level commands in subdirectories
1124- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1125
1126## 1.0.48
1127
1128- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1129- Added progress messages to Bash tool based on the last 5 lines of command output
1130- Added expanding variables support for MCP server configuration
1131- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1132- Improved IDE extension path handling when Claude Code runs in WSL
1133- Hooks: Added a PreCompact hook
1134- Vim mode: Added c, f/F, t/T
1135
1136## 1.0.45
1137
1138- Redesigned Search (Grep) tool with new tool input parameters and features
1139- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1140- Fixed config file corruption issue by enforcing atomic writes
1141- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1142- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1143- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1144
1145## 1.0.44
1146
1147- New /export command lets you quickly export a conversation for sharing
1148- MCP: resource_link tool results are now supported
1149- MCP: tool annotations and tool titles now display in /mcp view
1150- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1151
1152## 1.0.43
1153
1154- Fixed a bug where the theme selector was saving excessively
1155- Hooks: Added EPIPE system error handling
1156
1157## 1.0.42
1158
1159- Added tilde (`~`) expansion support to `/add-dir` command
1160
1161## 1.0.41
1162
1163- Hooks: Split Stop hook triggering into Stop and SubagentStop
1164- Hooks: Enabled optional timeout configuration for each command
1165- Hooks: Added "hook_event_name" to hook input
1166- Fixed a bug where MCP tools would display twice in tool list
1167- New tool parameters JSON for Bash tool in `tool_decision` event
1168
1169## 1.0.40
1170
1171- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1172
1173## 1.0.39
1174
1175- New Active Time metric in OpenTelemetry logging
1176
1177## 1.0.38
1178
1179- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1180
1181## 1.0.37
1182
1183- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1184
1185## 1.0.36
1186
1187- Web search now takes today's date into context
1188- Fixed a bug where stdio MCP servers were not terminating properly on exit
1189
1190## 1.0.35
1191
1192- Added support for MCP OAuth Authorization Server discovery
1193
1194## 1.0.34
1195
1196- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1197
1198## 1.0.33
1199
1200- Improved logging functionality with session ID support
1201- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1202- Improvements to plan mode
1203
1204## 1.0.32
1205
1206- Updated loopback config for litellm
1207- Added forceLoginMethod setting to bypass login selection screen
1208
1209## 1.0.31
1210
1211- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1212
1213## 1.0.30
1214
1215- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1216- Improved file path autocomplete with filename matching
1217- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1218- Enhanced jq regex support for complex filters with pipes and select
1219
1220## 1.0.29
1221
1222- Improved CJK character support in cursor navigation and rendering
1223
1224## 1.0.28
1225
1226- Slash commands: Fix selector display during history navigation
1227- Resizes images before upload to prevent API size limit errors
1228- Added XDG_CONFIG_HOME support to configuration directory
1229- Performance optimizations for memory usage
1230- New attributes (terminal.type, language) in OpenTelemetry logging
1231
1232## 1.0.27
1233
1234- Streamable HTTP MCP servers are now supported
1235- Remote MCP servers (SSE and HTTP) now support OAuth
1236- MCP resources can now be @-mentioned
1237- /resume slash command to switch conversations within Claude Code
1238
1239## 1.0.25
1240
1241- Slash commands: moved "project" and "user" prefixes to descriptions
1242- Slash commands: improved reliability for command discovery
1243- Improved support for Ghostty
1244- Improved web search reliability
1245
1246## 1.0.24
1247
1248- Improved /mcp output
1249- Fixed a bug where settings arrays got overwritten instead of merged
1250
1251## 1.0.23
1252
1253- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1254- Released Python SDK: pip install claude-code-sdk to get started
1255
1256## 1.0.22
1257
1258- SDK: Renamed `total_cost` to `total_cost_usd`
1259
1260## 1.0.21
1261
1262- Improved editing of files with tab-based indentation
1263- Fix for tool_use without matching tool_result errors
1264- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1265
1266## 1.0.18
1267
1268- Added --add-dir CLI argument for specifying additional working directories
1269- Added streaming input support without require -p flag
1270- Improved startup performance and session storage performance
1271- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1272- Added detailed MCP server tools display (/mcp)
1273- MCP authentication and permission improvements
1274- Added auto-reconnection for MCP SSE connections on disconnect
1275- Fixed issue where pasted content was lost when dialogs appeared
1276
1277## 1.0.17
1278
1279- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1280- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1281- MCP server list UI improvements
1282- Update Claude Code process title to display "claude" instead of "node"
1283
1284## 1.0.11
1285
1286- Claude Code can now also be used with a Claude Pro subscription
1287- Added /upgrade for smoother switching to Claude Max plans
1288- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1289- Improved shell configuration error handling
1290- Improved todo list handling during compaction
1291
1292## 1.0.10
1293
1294- Added markdown table support
1295- Improved streaming performance
1296
1297## 1.0.8
1298
1299- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1300- Increased default otel interval from 1s -> 5s
1301- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1302- Fixed a regression where search tools unnecessarily asked for permissions
1303- Added support for triggering thinking non-English languages
1304- Improved compacting UI
1305
1306## 1.0.7
1307
1308- Renamed /allowed-tools -> /permissions
1309- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1310- Deprecated claude config commands in favor of editing settings.json
1311- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1312- Improved error handling for /install-github-app
1313- Bugfixes, UI polish, and tool reliability improvements
1314
1315## 1.0.6
1316
1317- Improved edit reliability for tab-indented files
1318- Respect CLAUDE_CONFIG_DIR everywhere
1319- Reduced unnecessary tool permission prompts
1320- Added support for symlinks in @file typeahead
1321- Bugfixes, UI polish, and tool reliability improvements
1322
1323## 1.0.4
1324
1325- Fixed a bug where MCP tool errors weren't being parsed correctly
1326
1327## 1.0.1
1328
1329- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1330- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1331- Updated documentation links and OAuth process descriptions
1332
1333## 1.0.0
1334
1335- Claude Code is now generally available
1336- Introducing Sonnet 4 and Opus 4 models
1337
1338## 0.2.125
1339
1340- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1341- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1342
1343## 0.2.117
1344
1345- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1346- Introduced settings.cleanupPeriodDays
1347- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1348- Introduced --debug mode
1349
1350## 0.2.108
1351
1352- You can now send messages to Claude while it works to steer Claude in real-time
1353- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1354- Fixed a bug where thinking was not working in -p mode
1355- Fixed a regression in /cost reporting
1356- Deprecated MCP wizard interface in favor of other MCP commands
1357- Lots of other bugfixes and improvements
1358
1359## 0.2.107
1360
1361- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1362
1363## 0.2.106
1364
1365- MCP SSE server configs can now specify custom headers
1366- Fixed a bug where MCP permission prompt didn't always show correctly
1367
1368## 0.2.105
1369
1370- Claude can now search the web
1371- Moved system & account status to /status
1372- Added word movement keybindings for Vim
1373- Improved latency for startup, todo tool, and file edits
1374
1375## 0.2.102
1376
1377- Improved thinking triggering reliability
1378- Improved @mention reliability for images and folders
1379- You can now paste multiple large chunks into one prompt
1380
1381## 0.2.100
1382
1383- Fixed a crash caused by a stack overflow error
1384- Made db storage optional; missing db support disables --continue and --resume
1385
1386## 0.2.98
1387
1388- Fixed an issue where auto-compact was running twice
1389
1390## 0.2.96
1391
1392- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1393
1394## 0.2.93
1395
1396- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1397- Claude now has access to a Todo list that helps it stay on track and be more organized
1398
1399## 0.2.82
1400
1401- Added support for --disallowedTools
1402- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1403
1404## 0.2.75
1405
1406- Hit Enter to queue up additional messages while Claude is working
1407- Drag in or copy/paste image files directly into the prompt
1408- @-mention files to directly add them to context
1409- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1410- Improved performance for filename auto-complete
1411
1412## 0.2.74
1413
1414- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1415- Task tool can now perform writes and run bash commands
1416
1417## 0.2.72
1418
1419- Updated spinner to indicate tokens loaded and tool usage
1420
1421## 0.2.70
1422
1423- Network commands like curl are now available for Claude to use
1424- Claude can now run multiple web queries in parallel
1425- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1426
1427## 0.2.69
1428
1429- Fixed UI glitches with improved Select component behavior
1430- Enhanced terminal output display with better text truncation logic
1431
1432## 0.2.67
1433
1434- Shared project permission rules can be saved in .claude/settings.json
1435
1436## 0.2.66
1437
1438- Print mode (-p) now supports streaming output via --output-format=stream-json
1439- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1440
1441## 0.2.63
1442
1443- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1444
1445## 0.2.61
1446
1447- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1448- Enhanced image detection for more reliable clipboard paste functionality
1449- Fixed an issue where ESC key could crash the conversation history selector
1450
1451## 0.2.59
1452
1453- Copy+paste images directly into your prompt
1454- Improved progress indicators for bash and fetch tools
1455- Bugfixes for non-interactive mode (-p)
1456
1457## 0.2.54
1458
1459- Quickly add to Memory by starting your message with '#'
1460- Press ctrl+r to see full output for long tool results
1461- Added support for MCP SSE transport
1462
1463## 0.2.53
1464
1465- New web fetch tool lets Claude view URLs that you paste in
1466- Fixed a bug with JPEG detection
1467
1468## 0.2.50
1469
1470- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1471
1472## 0.2.49
1473
1474- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1475
1476## 0.2.47
1477
1478- Press Tab to auto-complete file and folder names
1479- Press Shift + Tab to toggle auto-accept for file edits
1480- Automatic conversation compaction for infinite conversation length (toggle with /config)
1481
1482## 0.2.44
1483
1484- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1485
1486## 0.2.41
1487
1488- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1489- MCP server startup no longer blocks the app from starting up
1490
1491## 0.2.37
1492
1493- New /release-notes command lets you view release notes at any time
1494- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1495
1496## 0.2.36
1497
1498- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1499- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1500
1501## 0.2.34
1502
1503- Vim bindings for text input - enable with /vim or /config
1504
1505## 0.2.32
1506
1507- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1508- Fix for some PersistentShell issues
1509
1510## 0.2.31
1511
1512- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1513- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1514
1515## 0.2.30
1516
1517- Added ANSI color theme for better terminal compatibility
1518- Fixed issue where slash command arguments weren't being sent properly
1519- (Mac-only) API keys are now stored in macOS Keychain
1520
1521## 0.2.26
1522
1523- New /approved-tools command for managing tool permissions
1524- Word-level diff display for improved code readability
1525- Fuzzy matching for slash commands
1526
1527## 0.2.21
1528
1529- Fuzzy matching for /commands