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 bd8aff2 · on Sep 26, 2025 History
706 lines · 25.0 KBPreviewCodeRaw
1# Changelog
2
3## 1.0.126
4
5- Enable /context command for Bedrock and Vertex
6- Add mTLS support for HTTP-based OpenTelemetry exporters
7
8## 1.0.124
9
10- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
11- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
12- No longer inform Claude of the list of allowed tools when permission is denied
13- Fixed security vulnerability in Bash tool permission checks
14- Improved VSCode extension performance for large files
15
16## 1.0.123
17
18- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
19- Fixed thinking mode triggering on negation phrases like "don't think"
20- Fixed rendering performance degradation during token streaming
21- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
22- Enhanced BashTool environment snapshot logging
23- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
24- Migrated --debug logging to a file, to enable easy tailing & filtering
25
26## 1.0.120
27
28- Fix input lag during typing, especially noticeable with large prompts
29- Improved VSCode extension command registry and sessions dialog user experience
30- Enhanced sessions dialog responsiveness and visual feedback
31- Fixed IDE compatibility issue by removing worktree support check
32- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
33
34## 1.0.119
35
36- Fix Windows issue where process visually freezes on entering interactive mode
37- Support dynamic headers for MCP servers via headersHelper configuration
38- Fix thinking mode not working in headless sessions
39- Fix slash commands now properly update allowed tools instead of replacing them
40
41## 1.0.117
42
43- Add Ctrl-R history search to recall previous commands like bash/zsh
44- Fix input lag while typing, especially on Windows
45- Add sed command to auto-allowed commands in acceptEdits mode
46- Fix Windows PATH comparison to be case-insensitive for drive letters
47- Add permissions management hint to /add-dir output
48
49## 1.0.115
50
51- Improve thinking mode display with enhanced visual effects
52- Type /t to temporarily disable thinking mode in your prompt
53- Improve path validation for glob and grep tools
54- Show condensed output for post-tool hooks to reduce visual clutter
55- Fix visual feedback when loading state completes
56- Improve UI consistency for permission request dialogs
57
58## 1.0.113
59
60- Deprecated piped input in interactive mode
61- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
62
63## 1.0.112
64
65- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
66- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
67- Hooks: Added systemMessage support for SessionEnd hooks
68- Added `spinnerTipsEnabled` setting to disable spinner tips
69- IDE: Various improvements and bug fixes
70
71## 1.0.111
72
73- /model now validates provided model names
74- Fixed Bash tool crashes caused by malformed shell syntax parsing
75
76## 1.0.110
77
78- /terminal-setup command now supports WezTerm
79- MCP: OAuth tokens now proactively refresh before expiration
80- Fixed reliability issues with background Bash processes
81
82## 1.0.109
83
84- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
85
86## 1.0.106
87
88- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
89
90## 1.0.97
91
92- Settings: /doctor now validates permission rule syntax and suggests corrections
93
94## 1.0.94
95
96- Vertex: add support for global endpoints for supported models
97- /memory command now allows direct editing of all imported memory files
98- SDK: Add custom tools as callbacks
99- Added /todos command to list current todo items
100
101## 1.0.93
102
103- Windows: Add alt + v shortcut for pasting images from clipboard
104- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
105
106## 1.0.90
107
108- Settings file changes take effect immediately - no restart required
109
110## 1.0.88
111
112- Fixed issue causing "OAuth authentication is currently not supported"
113- Status line input now includes `exceeds_200k_tokens`
114- Fixed incorrect usage tracking in /cost.
115- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
116- Bedrock: Updated default Sonnet model to Sonnet 4
117
118## 1.0.86
119
120- Added /context to help users self-serve debug context issues
121- SDK: Added UUID support for all SDK messages
122- SDK: Added `--replay-user-messages` to replay user messages back to stdout
123
124## 1.0.85
125
126- Status line input now includes session cost info
127- Hooks: Introduced SessionEnd hook
128
129## 1.0.84
130
131- Fix tool_use/tool_result id mismatch error when network is unstable
132- Fix Claude sometimes ignoring real-time steering when wrapping up a task
133- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
134- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
135
136## 1.0.83
137
138- @-mention: Support files with spaces in path
139- New shimmering spinner
140
141## 1.0.82
142
143- SDK: Add request cancellation support
144- SDK: New additionalDirectories option to search custom paths, improved slash command processing
145- Settings: Validation prevents invalid fields in .claude/settings.json files
146- MCP: Improve tool name consistency
147- Bash: Fix crash when Claude tries to automatically read large files
148
149## 1.0.81
150
151- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
152- Agents: Fix custom agent loading when agent files are unparsable
153
154## 1.0.80
155
156- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
157
158## 1.0.77
159
160- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
161- SDK: Add session support and permission denial tracking
162- Fix token limit errors in conversation summarization
163- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
164
165## 1.0.73
166
167- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
168- MCP: Press Esc to cancel OAuth authentication flows
169- Bash: Improved command validation and reduced false security warnings
170- UI: Enhanced spinner animations and status line visual hierarchy
171- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
172
173## 1.0.72
174
175- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
176
177## 1.0.71
178
179- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
180- Customizable status line: add your terminal prompt to Claude Code with /statusline
181
182## 1.0.70
183
184- Performance: Optimized message rendering for better performance with large contexts
185- Windows: Fixed native file search, ripgrep, and subagent functionality
186- Added support for @-mentions in slash command arguments
187
188## 1.0.69
189
190- Upgraded Opus to version 4.1
191
192## 1.0.68
193
194- Fix incorrect model names being used for certain commands like `/pr-comments`
195- 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.
196- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
197- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
198- SDK: Added canUseTool callback support for tool confirmation
199- Added `disableAllHooks` setting
200- Improved file suggestions performance in large repos
201
202## 1.0.65
203
204- IDE: Fixed connection stability issues and error handling for diagnostics
205- Windows: Fixed shell environment setup for users without .bashrc files
206
207## 1.0.64
208
209- Agents: Added model customization support - you can now specify which model an agent should use
210- Agents: Fixed unintended access to the recursive agent tool
211- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
212- SDK: Fixed user input tracking across multi-turn conversations
213- Added hidden files to file search and @-mention suggestions
214
215## 1.0.63
216
217- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
218
219## 1.0.62
220
221- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
222- Hooks: Added SessionStart hook for new session initialization
223- /add-dir command now supports typeahead for directory paths
224- Improved network connectivity check reliability
225
226## 1.0.61
227
228- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
229- Settings: Added `--settings` flag to load settings from a JSON file
230- Settings: Fixed resolution of settings files paths that are symlinks
231- OTEL: Fixed reporting of wrong organization after authentication changes
232- Slash commands: Fixed permissions checking for allowed-tools with Bash
233- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
234- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
235- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
236
237## 1.0.60
238
239- You can now create custom subagents for specialized tasks! Run /agents to get started
240
241## 1.0.59
242
243- SDK: Added tool confirmation support with canUseTool callback
244- SDK: Allow specifying env for spawned process
245- Hooks: Exposed PermissionDecision to hooks (including "ask")
246- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
247- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
248
249## 1.0.58
250
251- Added support for reading PDFs
252- MCP: Improved server health status display in 'claude mcp list'
253- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
254
255## 1.0.57
256
257- Added support for specifying a model in slash commands
258- Improved permission messages to help Claude understand allowed tools
259- Fix: Remove trailing newlines from bash output in terminal wrapping
260
261## 1.0.56
262
263- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
264- Fixes for WSL IDE detection
265- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
266
267## 1.0.55
268
269- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
270- Windows: fixed Ctrl+Z crash
271- SDK: Added ability to capture error logging
272- Add --system-prompt-file option to override system prompt in print mode
273
274## 1.0.54
275
276- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
277- Custom slash commands: Added argument-hint to frontmatter
278- Windows: OAuth uses port 45454 and properly constructs browser URL
279- Windows: mode switching now uses alt + m, and plan mode renders properly
280- Shell: Switch to in-memory shell snapshot to fix file-related errors
281
282## 1.0.53
283
284- Updated @-mention file truncation from 100 lines to 2000 lines
285- 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).
286
287## 1.0.52
288
289- Added support for MCP server instructions
290
291## 1.0.51
292
293- Added support for native Windows (requires Git for Windows)
294- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
295- Settings: /doctor can now help you identify and fix invalid setting files
296- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
297- Increased auto-compact warning threshold from 60% to 80%
298- Fixed an issue with handling user directories with spaces for shell snapshots
299- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
300- Custom slash commands: Fixed user-level commands in subdirectories
301- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
302
303## 1.0.48
304
305- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
306- Added progress messages to Bash tool based on the last 5 lines of command output
307- Added expanding variables support for MCP server configuration
308- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
309- Improved IDE extension path handling when Claude Code runs in WSL
310- Hooks: Added a PreCompact hook
311- Vim mode: Added c, f/F, t/T
312
313## 1.0.45
314
315- Redesigned Search (Grep) tool with new tool input parameters and features
316- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
317- Fixed config file corruption issue by enforcing atomic writes
318- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
319- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
320- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
321
322## 1.0.44
323
324- New /export command lets you quickly export a conversation for sharing
325- MCP: resource_link tool results are now supported
326- MCP: tool annotations and tool titles now display in /mcp view
327- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
328
329## 1.0.43
330
331- Fixed a bug where the theme selector was saving excessively
332- Hooks: Added EPIPE system error handling
333
334## 1.0.42
335
336- Added tilde (`~`) expansion support to `/add-dir` command
337
338## 1.0.41
339
340- Hooks: Split Stop hook triggering into Stop and SubagentStop
341- Hooks: Enabled optional timeout configuration for each command
342- Hooks: Added "hook_event_name" to hook input
343- Fixed a bug where MCP tools would display twice in tool list
344- New tool parameters JSON for Bash tool in `tool_decision` event
345
346## 1.0.40
347
348- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
349
350## 1.0.39
351
352- New Active Time metric in OpenTelemetry logging
353
354## 1.0.38
355
356- 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
357
358## 1.0.37
359
360- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
361
362## 1.0.36
363
364- Web search now takes today's date into context
365- Fixed a bug where stdio MCP servers were not terminating properly on exit
366
367## 1.0.35
368
369- Added support for MCP OAuth Authorization Server discovery
370
371## 1.0.34
372
373- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
374
375## 1.0.33
376
377- Improved logging functionality with session ID support
378- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
379- Improvements to plan mode
380
381## 1.0.32
382
383- Updated loopback config for litellm
384- Added forceLoginMethod setting to bypass login selection screen
385
386## 1.0.31
387
388- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
389
390## 1.0.30
391
392- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
393- Improved file path autocomplete with filename matching
394- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
395- Enhanced jq regex support for complex filters with pipes and select
396
397## 1.0.29
398
399- Improved CJK character support in cursor navigation and rendering
400
401## 1.0.28
402
403- Slash commands: Fix selector display during history navigation
404- Resizes images before upload to prevent API size limit errors
405- Added XDG_CONFIG_HOME support to configuration directory
406- Performance optimizations for memory usage
407- New attributes (terminal.type, language) in OpenTelemetry logging
408
409## 1.0.27
410
411- Streamable HTTP MCP servers are now supported
412- Remote MCP servers (SSE and HTTP) now support OAuth
413- MCP resources can now be @-mentioned
414- /resume slash command to switch conversations within Claude Code
415
416## 1.0.25
417
418- Slash commands: moved "project" and "user" prefixes to descriptions
419- Slash commands: improved reliability for command discovery
420- Improved support for Ghostty
421- Improved web search reliability
422
423## 1.0.24
424
425- Improved /mcp output
426- Fixed a bug where settings arrays got overwritten instead of merged
427
428## 1.0.23
429
430- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
431- Released Python SDK: pip install claude-code-sdk to get started
432
433## 1.0.22
434
435- SDK: Renamed `total_cost` to `total_cost_usd`
436
437## 1.0.21
438
439- Improved editing of files with tab-based indentation
440- Fix for tool_use without matching tool_result errors
441- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
442
443## 1.0.18
444
445- Added --add-dir CLI argument for specifying additional working directories
446- Added streaming input support without require -p flag
447- Improved startup performance and session storage performance
448- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
449- Added detailed MCP server tools display (/mcp)
450- MCP authentication and permission improvements
451- Added auto-reconnection for MCP SSE connections on disconnect
452- Fixed issue where pasted content was lost when dialogs appeared
453
454## 1.0.17
455
456- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
457- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
458- MCP server list UI improvements
459- Update Claude Code process title to display "claude" instead of "node"
460
461## 1.0.11
462
463- Claude Code can now also be used with a Claude Pro subscription
464- Added /upgrade for smoother switching to Claude Max plans
465- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
466- Improved shell configuration error handling
467- Improved todo list handling during compaction
468
469## 1.0.10
470
471- Added markdown table support
472- Improved streaming performance
473
474## 1.0.8
475
476- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
477- Increased default otel interval from 1s -> 5s
478- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
479- Fixed a regression where search tools unnecessarily asked for permissions
480- Added support for triggering thinking non-English languages
481- Improved compacting UI
482
483## 1.0.7
484
485- Renamed /allowed-tools -> /permissions
486- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
487- Deprecated claude config commands in favor of editing settings.json
488- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
489- Improved error handling for /install-github-app
490- Bugfixes, UI polish, and tool reliability improvements
491
492## 1.0.6
493
494- Improved edit reliability for tab-indented files
495- Respect CLAUDE_CONFIG_DIR everywhere
496- Reduced unnecessary tool permission prompts
497- Added support for symlinks in @file typeahead
498- Bugfixes, UI polish, and tool reliability improvements
499
500## 1.0.4
501
502- Fixed a bug where MCP tool errors weren't being parsed correctly
503
504## 1.0.1
505
506- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
507- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
508- Updated documentation links and OAuth process descriptions
509
510## 1.0.0
511
512- Claude Code is now generally available
513- Introducing Sonnet 4 and Opus 4 models
514
515## 0.2.125
516
517- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
518- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
519
520## 0.2.117
521
522- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
523- Introduced settings.cleanupPeriodDays
524- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
525- Introduced --debug mode
526
527## 0.2.108
528
529- You can now send messages to Claude while it works to steer Claude in real-time
530- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
531- Fixed a bug where thinking was not working in -p mode
532- Fixed a regression in /cost reporting
533- Deprecated MCP wizard interface in favor of other MCP commands
534- Lots of other bugfixes and improvements
535
536## 0.2.107
537
538- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
539
540## 0.2.106
541
542- MCP SSE server configs can now specify custom headers
543- Fixed a bug where MCP permission prompt didn't always show correctly
544
545## 0.2.105
546
547- Claude can now search the web
548- Moved system & account status to /status
549- Added word movement keybindings for Vim
550- Improved latency for startup, todo tool, and file edits
551
552## 0.2.102
553
554- Improved thinking triggering reliability
555- Improved @mention reliability for images and folders
556- You can now paste multiple large chunks into one prompt
557
558## 0.2.100
559
560- Fixed a crash caused by a stack overflow error
561- Made db storage optional; missing db support disables --continue and --resume
562
563## 0.2.98
564
565- Fixed an issue where auto-compact was running twice
566
567## 0.2.96
568
569- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
570
571## 0.2.93
572
573- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
574- Claude now has access to a Todo list that helps it stay on track and be more organized
575
576## 0.2.82
577
578- Added support for --disallowedTools
579- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
580
581## 0.2.75
582
583- Hit Enter to queue up additional messages while Claude is working
584- Drag in or copy/paste image files directly into the prompt
585- @-mention files to directly add them to context
586- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
587- Improved performance for filename auto-complete
588
589## 0.2.74
590
591- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
592- Task tool can now perform writes and run bash commands
593
594## 0.2.72
595
596- Updated spinner to indicate tokens loaded and tool usage
597
598## 0.2.70
599
600- Network commands like curl are now available for Claude to use
601- Claude can now run multiple web queries in parallel
602- Pressing ESC once immediately interrupts Claude in Auto-accept mode
603
604## 0.2.69
605
606- Fixed UI glitches with improved Select component behavior
607- Enhanced terminal output display with better text truncation logic
608
609## 0.2.67
610
611- Shared project permission rules can be saved in .claude/settings.json
612
613## 0.2.66
614
615- Print mode (-p) now supports streaming output via --output-format=stream-json
616- Fixed issue where pasting could trigger memory or bash mode unexpectedly
617
618## 0.2.63
619
620- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
621
622## 0.2.61
623
624- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
625- Enhanced image detection for more reliable clipboard paste functionality
626- Fixed an issue where ESC key could crash the conversation history selector
627
628## 0.2.59
629
630- Copy+paste images directly into your prompt
631- Improved progress indicators for bash and fetch tools
632- Bugfixes for non-interactive mode (-p)
633
634## 0.2.54
635
636- Quickly add to Memory by starting your message with '#'
637- Press ctrl+r to see full output for long tool results
638- Added support for MCP SSE transport
639
640## 0.2.53
641
642- New web fetch tool lets Claude view URLs that you paste in
643- Fixed a bug with JPEG detection
644
645## 0.2.50
646
647- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
648
649## 0.2.49
650
651- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
652
653## 0.2.47
654
655- Press Tab to auto-complete file and folder names
656- Press Shift + Tab to toggle auto-accept for file edits
657- Automatic conversation compaction for infinite conversation length (toggle with /config)
658
659## 0.2.44
660
661- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
662
663## 0.2.41
664
665- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
666- MCP server startup no longer blocks the app from starting up
667
668## 0.2.37
669
670- New /release-notes command lets you view release notes at any time
671- `claude config add/remove` commands now accept multiple values separated by commas or spaces
672
673## 0.2.36
674
675- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
676- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
677
678## 0.2.34
679
680- Vim bindings for text input - enable with /vim or /config
681
682## 0.2.32
683
684- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
685- Fix for some PersistentShell issues
686
687## 0.2.31
688
689- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
690- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
691
692## 0.2.30
693
694- Added ANSI color theme for better terminal compatibility
695- Fixed issue where slash command arguments weren't being sent properly
696- (Mac-only) API keys are now stored in macOS Keychain
697
698## 0.2.26
699
700- New /approved-tools command for managing tool permissions
701- Word-level diff display for improved code readability
702- Fuzzy matching for slash commands
703
704## 0.2.21
705
706- Fuzzy matching for /commands