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 f5f6c76 · on Sep 23, 2025 History
693 lines · 24.5 KBPreviewCodeRaw
1# Changelog
2
3## 1.0.123
4
5- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
6- Fixed thinking mode triggering on negation phrases like "don't think"
7- Fixed rendering performance degradation during token streaming
8- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
9- Enhanced BashTool environment snapshot logging
10- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
11- Migrated --debug logging to a file, to enable easy tailing & filtering
12
13## 1.0.120
14
15- Fix input lag during typing, especially noticeable with large prompts
16- Improved VSCode extension command registry and sessions dialog user experience
17- Enhanced sessions dialog responsiveness and visual feedback
18- Fixed IDE compatibility issue by removing worktree support check
19- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
20
21## 1.0.119
22
23- Fix Windows issue where process visually freezes on entering interactive mode
24- Support dynamic headers for MCP servers via headersHelper configuration
25- Fix thinking mode not working in headless sessions
26- Fix slash commands now properly update allowed tools instead of replacing them
27
28## 1.0.117
29
30- Add Ctrl-R history search to recall previous commands like bash/zsh
31- Fix input lag while typing, especially on Windows
32- Add sed command to auto-allowed commands in acceptEdits mode
33- Fix Windows PATH comparison to be case-insensitive for drive letters
34- Add permissions management hint to /add-dir output
35
36## 1.0.115
37
38- Improve thinking mode display with enhanced visual effects
39- Type /t to temporarily disable thinking mode in your prompt
40- Improve path validation for glob and grep tools
41- Show condensed output for post-tool hooks to reduce visual clutter
42- Fix visual feedback when loading state completes
43- Improve UI consistency for permission request dialogs
44
45## 1.0.113
46
47- Deprecated piped input in interactive mode
48- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
49
50## 1.0.112
51
52- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
53- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
54- Hooks: Added systemMessage support for SessionEnd hooks
55- Added `spinnerTipsEnabled` setting to disable spinner tips
56- IDE: Various improvements and bug fixes
57
58## 1.0.111
59
60- /model now validates provided model names
61- Fixed Bash tool crashes caused by malformed shell syntax parsing
62
63## 1.0.110
64
65- /terminal-setup command now supports WezTerm
66- MCP: OAuth tokens now proactively refresh before expiration
67- Fixed reliability issues with background Bash processes
68
69## 1.0.109
70
71- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
72
73## 1.0.106
74
75- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
76
77## 1.0.97
78
79- Settings: /doctor now validates permission rule syntax and suggests corrections
80
81## 1.0.94
82
83- Vertex: add support for global endpoints for supported models
84- /memory command now allows direct editing of all imported memory files
85- SDK: Add custom tools as callbacks
86- Added /todos command to list current todo items
87
88## 1.0.93
89
90- Windows: Add alt + v shortcut for pasting images from clipboard
91- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
92
93## 1.0.90
94
95- Settings file changes take effect immediately - no restart required
96
97## 1.0.88
98
99- Fixed issue causing "OAuth authentication is currently not supported"
100- Status line input now includes `exceeds_200k_tokens`
101- Fixed incorrect usage tracking in /cost.
102- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
103- Bedrock: Updated default Sonnet model to Sonnet 4
104
105## 1.0.86
106
107- Added /context to help users self-serve debug context issues
108- SDK: Added UUID support for all SDK messages
109- SDK: Added `--replay-user-messages` to replay user messages back to stdout
110
111## 1.0.85
112
113- Status line input now includes session cost info
114- Hooks: Introduced SessionEnd hook
115
116## 1.0.84
117
118- Fix tool_use/tool_result id mismatch error when network is unstable
119- Fix Claude sometimes ignoring real-time steering when wrapping up a task
120- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
121- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
122
123## 1.0.83
124
125- @-mention: Support files with spaces in path
126- New shimmering spinner
127
128## 1.0.82
129
130- SDK: Add request cancellation support
131- SDK: New additionalDirectories option to search custom paths, improved slash command processing
132- Settings: Validation prevents invalid fields in .claude/settings.json files
133- MCP: Improve tool name consistency
134- Bash: Fix crash when Claude tries to automatically read large files
135
136## 1.0.81
137
138- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
139- Agents: Fix custom agent loading when agent files are unparsable
140
141## 1.0.80
142
143- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
144
145## 1.0.77
146
147- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
148- SDK: Add session support and permission denial tracking
149- Fix token limit errors in conversation summarization
150- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
151
152## 1.0.73
153
154- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
155- MCP: Press Esc to cancel OAuth authentication flows
156- Bash: Improved command validation and reduced false security warnings
157- UI: Enhanced spinner animations and status line visual hierarchy
158- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
159
160## 1.0.72
161
162- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
163
164## 1.0.71
165
166- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
167- Customizable status line: add your terminal prompt to Claude Code with /statusline
168
169## 1.0.70
170
171- Performance: Optimized message rendering for better performance with large contexts
172- Windows: Fixed native file search, ripgrep, and subagent functionality
173- Added support for @-mentions in slash command arguments
174
175## 1.0.69
176
177- Upgraded Opus to version 4.1
178
179## 1.0.68
180
181- Fix incorrect model names being used for certain commands like `/pr-comments`
182- 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.
183- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
184- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
185- SDK: Added canUseTool callback support for tool confirmation
186- Added `disableAllHooks` setting
187- Improved file suggestions performance in large repos
188
189## 1.0.65
190
191- IDE: Fixed connection stability issues and error handling for diagnostics
192- Windows: Fixed shell environment setup for users without .bashrc files
193
194## 1.0.64
195
196- Agents: Added model customization support - you can now specify which model an agent should use
197- Agents: Fixed unintended access to the recursive agent tool
198- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
199- SDK: Fixed user input tracking across multi-turn conversations
200- Added hidden files to file search and @-mention suggestions
201
202## 1.0.63
203
204- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
205
206## 1.0.62
207
208- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
209- Hooks: Added SessionStart hook for new session initialization
210- /add-dir command now supports typeahead for directory paths
211- Improved network connectivity check reliability
212
213## 1.0.61
214
215- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
216- Settings: Added `--settings` flag to load settings from a JSON file
217- Settings: Fixed resolution of settings files paths that are symlinks
218- OTEL: Fixed reporting of wrong organization after authentication changes
219- Slash commands: Fixed permissions checking for allowed-tools with Bash
220- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
221- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
222- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
223
224## 1.0.60
225
226- You can now create custom subagents for specialized tasks! Run /agents to get started
227
228## 1.0.59
229
230- SDK: Added tool confirmation support with canUseTool callback
231- SDK: Allow specifying env for spawned process
232- Hooks: Exposed PermissionDecision to hooks (including "ask")
233- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
234- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
235
236## 1.0.58
237
238- Added support for reading PDFs
239- MCP: Improved server health status display in 'claude mcp list'
240- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
241
242## 1.0.57
243
244- Added support for specifying a model in slash commands
245- Improved permission messages to help Claude understand allowed tools
246- Fix: Remove trailing newlines from bash output in terminal wrapping
247
248## 1.0.56
249
250- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
251- Fixes for WSL IDE detection
252- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
253
254## 1.0.55
255
256- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
257- Windows: fixed Ctrl+Z crash
258- SDK: Added ability to capture error logging
259- Add --system-prompt-file option to override system prompt in print mode
260
261## 1.0.54
262
263- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
264- Custom slash commands: Added argument-hint to frontmatter
265- Windows: OAuth uses port 45454 and properly constructs browser URL
266- Windows: mode switching now uses alt + m, and plan mode renders properly
267- Shell: Switch to in-memory shell snapshot to fix file-related errors
268
269## 1.0.53
270
271- Updated @-mention file truncation from 100 lines to 2000 lines
272- 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).
273
274## 1.0.52
275
276- Added support for MCP server instructions
277
278## 1.0.51
279
280- Added support for native Windows (requires Git for Windows)
281- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
282- Settings: /doctor can now help you identify and fix invalid setting files
283- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
284- Increased auto-compact warning threshold from 60% to 80%
285- Fixed an issue with handling user directories with spaces for shell snapshots
286- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
287- Custom slash commands: Fixed user-level commands in subdirectories
288- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
289
290## 1.0.48
291
292- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
293- Added progress messages to Bash tool based on the last 5 lines of command output
294- Added expanding variables support for MCP server configuration
295- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
296- Improved IDE extension path handling when Claude Code runs in WSL
297- Hooks: Added a PreCompact hook
298- Vim mode: Added c, f/F, t/T
299
300## 1.0.45
301
302- Redesigned Search (Grep) tool with new tool input parameters and features
303- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
304- Fixed config file corruption issue by enforcing atomic writes
305- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
306- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
307- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
308
309## 1.0.44
310
311- New /export command lets you quickly export a conversation for sharing
312- MCP: resource_link tool results are now supported
313- MCP: tool annotations and tool titles now display in /mcp view
314- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
315
316## 1.0.43
317
318- Fixed a bug where the theme selector was saving excessively
319- Hooks: Added EPIPE system error handling
320
321## 1.0.42
322
323- Added tilde (`~`) expansion support to `/add-dir` command
324
325## 1.0.41
326
327- Hooks: Split Stop hook triggering into Stop and SubagentStop
328- Hooks: Enabled optional timeout configuration for each command
329- Hooks: Added "hook_event_name" to hook input
330- Fixed a bug where MCP tools would display twice in tool list
331- New tool parameters JSON for Bash tool in `tool_decision` event
332
333## 1.0.40
334
335- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
336
337## 1.0.39
338
339- New Active Time metric in OpenTelemetry logging
340
341## 1.0.38
342
343- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://docs.claude.com/en/docs/claude-code/hooks
344
345## 1.0.37
346
347- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
348
349## 1.0.36
350
351- Web search now takes today's date into context
352- Fixed a bug where stdio MCP servers were not terminating properly on exit
353
354## 1.0.35
355
356- Added support for MCP OAuth Authorization Server discovery
357
358## 1.0.34
359
360- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
361
362## 1.0.33
363
364- Improved logging functionality with session ID support
365- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
366- Improvements to plan mode
367
368## 1.0.32
369
370- Updated loopback config for litellm
371- Added forceLoginMethod setting to bypass login selection screen
372
373## 1.0.31
374
375- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
376
377## 1.0.30
378
379- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
380- Improved file path autocomplete with filename matching
381- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
382- Enhanced jq regex support for complex filters with pipes and select
383
384## 1.0.29
385
386- Improved CJK character support in cursor navigation and rendering
387
388## 1.0.28
389
390- Slash commands: Fix selector display during history navigation
391- Resizes images before upload to prevent API size limit errors
392- Added XDG_CONFIG_HOME support to configuration directory
393- Performance optimizations for memory usage
394- New attributes (terminal.type, language) in OpenTelemetry logging
395
396## 1.0.27
397
398- Streamable HTTP MCP servers are now supported
399- Remote MCP servers (SSE and HTTP) now support OAuth
400- MCP resources can now be @-mentioned
401- /resume slash command to switch conversations within Claude Code
402
403## 1.0.25
404
405- Slash commands: moved "project" and "user" prefixes to descriptions
406- Slash commands: improved reliability for command discovery
407- Improved support for Ghostty
408- Improved web search reliability
409
410## 1.0.24
411
412- Improved /mcp output
413- Fixed a bug where settings arrays got overwritten instead of merged
414
415## 1.0.23
416
417- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
418- Released Python SDK: pip install claude-code-sdk to get started
419
420## 1.0.22
421
422- SDK: Renamed `total_cost` to `total_cost_usd`
423
424## 1.0.21
425
426- Improved editing of files with tab-based indentation
427- Fix for tool_use without matching tool_result errors
428- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
429
430## 1.0.18
431
432- Added --add-dir CLI argument for specifying additional working directories
433- Added streaming input support without require -p flag
434- Improved startup performance and session storage performance
435- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
436- Added detailed MCP server tools display (/mcp)
437- MCP authentication and permission improvements
438- Added auto-reconnection for MCP SSE connections on disconnect
439- Fixed issue where pasted content was lost when dialogs appeared
440
441## 1.0.17
442
443- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
444- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
445- MCP server list UI improvements
446- Update Claude Code process title to display "claude" instead of "node"
447
448## 1.0.11
449
450- Claude Code can now also be used with a Claude Pro subscription
451- Added /upgrade for smoother switching to Claude Max plans
452- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
453- Improved shell configuration error handling
454- Improved todo list handling during compaction
455
456## 1.0.10
457
458- Added markdown table support
459- Improved streaming performance
460
461## 1.0.8
462
463- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
464- Increased default otel interval from 1s -> 5s
465- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
466- Fixed a regression where search tools unnecessarily asked for permissions
467- Added support for triggering thinking non-English languages
468- Improved compacting UI
469
470## 1.0.7
471
472- Renamed /allowed-tools -> /permissions
473- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
474- Deprecated claude config commands in favor of editing settings.json
475- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
476- Improved error handling for /install-github-app
477- Bugfixes, UI polish, and tool reliability improvements
478
479## 1.0.6
480
481- Improved edit reliability for tab-indented files
482- Respect CLAUDE_CONFIG_DIR everywhere
483- Reduced unnecessary tool permission prompts
484- Added support for symlinks in @file typeahead
485- Bugfixes, UI polish, and tool reliability improvements
486
487## 1.0.4
488
489- Fixed a bug where MCP tool errors weren't being parsed correctly
490
491## 1.0.1
492
493- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
494- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
495- Updated documentation links and OAuth process descriptions
496
497## 1.0.0
498
499- Claude Code is now generally available
500- Introducing Sonnet 4 and Opus 4 models
501
502## 0.2.125
503
504- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
505- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
506
507## 0.2.117
508
509- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
510- Introduced settings.cleanupPeriodDays
511- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
512- Introduced --debug mode
513
514## 0.2.108
515
516- You can now send messages to Claude while it works to steer Claude in real-time
517- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
518- Fixed a bug where thinking was not working in -p mode
519- Fixed a regression in /cost reporting
520- Deprecated MCP wizard interface in favor of other MCP commands
521- Lots of other bugfixes and improvements
522
523## 0.2.107
524
525- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
526
527## 0.2.106
528
529- MCP SSE server configs can now specify custom headers
530- Fixed a bug where MCP permission prompt didn't always show correctly
531
532## 0.2.105
533
534- Claude can now search the web
535- Moved system & account status to /status
536- Added word movement keybindings for Vim
537- Improved latency for startup, todo tool, and file edits
538
539## 0.2.102
540
541- Improved thinking triggering reliability
542- Improved @mention reliability for images and folders
543- You can now paste multiple large chunks into one prompt
544
545## 0.2.100
546
547- Fixed a crash caused by a stack overflow error
548- Made db storage optional; missing db support disables --continue and --resume
549
550## 0.2.98
551
552- Fixed an issue where auto-compact was running twice
553
554## 0.2.96
555
556- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
557
558## 0.2.93
559
560- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
561- Claude now has access to a Todo list that helps it stay on track and be more organized
562
563## 0.2.82
564
565- Added support for --disallowedTools
566- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
567
568## 0.2.75
569
570- Hit Enter to queue up additional messages while Claude is working
571- Drag in or copy/paste image files directly into the prompt
572- @-mention files to directly add them to context
573- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
574- Improved performance for filename auto-complete
575
576## 0.2.74
577
578- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
579- Task tool can now perform writes and run bash commands
580
581## 0.2.72
582
583- Updated spinner to indicate tokens loaded and tool usage
584
585## 0.2.70
586
587- Network commands like curl are now available for Claude to use
588- Claude can now run multiple web queries in parallel
589- Pressing ESC once immediately interrupts Claude in Auto-accept mode
590
591## 0.2.69
592
593- Fixed UI glitches with improved Select component behavior
594- Enhanced terminal output display with better text truncation logic
595
596## 0.2.67
597
598- Shared project permission rules can be saved in .claude/settings.json
599
600## 0.2.66
601
602- Print mode (-p) now supports streaming output via --output-format=stream-json
603- Fixed issue where pasting could trigger memory or bash mode unexpectedly
604
605## 0.2.63
606
607- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
608
609## 0.2.61
610
611- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
612- Enhanced image detection for more reliable clipboard paste functionality
613- Fixed an issue where ESC key could crash the conversation history selector
614
615## 0.2.59
616
617- Copy+paste images directly into your prompt
618- Improved progress indicators for bash and fetch tools
619- Bugfixes for non-interactive mode (-p)
620
621## 0.2.54
622
623- Quickly add to Memory by starting your message with '#'
624- Press ctrl+r to see full output for long tool results
625- Added support for MCP SSE transport
626
627## 0.2.53
628
629- New web fetch tool lets Claude view URLs that you paste in
630- Fixed a bug with JPEG detection
631
632## 0.2.50
633
634- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
635
636## 0.2.49
637
638- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
639
640## 0.2.47
641
642- Press Tab to auto-complete file and folder names
643- Press Shift + Tab to toggle auto-accept for file edits
644- Automatic conversation compaction for infinite conversation length (toggle with /config)
645
646## 0.2.44
647
648- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
649
650## 0.2.41
651
652- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
653- MCP server startup no longer blocks the app from starting up
654
655## 0.2.37
656
657- New /release-notes command lets you view release notes at any time
658- `claude config add/remove` commands now accept multiple values separated by commas or spaces
659
660## 0.2.36
661
662- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
663- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
664
665## 0.2.34
666
667- Vim bindings for text input - enable with /vim or /config
668
669## 0.2.32
670
671- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
672- Fix for some PersistentShell issues
673
674## 0.2.31
675
676- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
677- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
678
679## 0.2.30
680
681- Added ANSI color theme for better terminal compatibility
682- Fixed issue where slash command arguments weren't being sent properly
683- (Mac-only) API keys are now stored in macOS Keychain
684
685## 0.2.26
686
687- New /approved-tools command for managing tool permissions
688- Word-level diff display for improved code readability
689- Fuzzy matching for slash commands
690
691## 0.2.21
692
693- Fuzzy matching for /commands