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 1b41969 · on Oct 24, 2025 History
843 lines · 30.4 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.27
4
5- New UI for permission prompts
6- Added current branch filtering and search to session resume screen for easier navigation
7- Fixed directory @-mention causing "No assistant message found" error
8- VSCode Extension: Add config setting to include .gitignored files in file searches
9- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
10
11## 2.0.25
12
13- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://docs.claude.com/en/docs/claude-code/sdk/migration-guide
14
15## 2.0.24
16
17- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
18- Claude Code Web: Support for Web -> CLI teleport
19- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
20
21## 2.0.22
22
23- Fixed content layout shift when scrolling through slash commands
24- IDE: Add toggle to enable/disable thinking.
25- Fix bug causing duplicate permission prompts with parallel tool calls
26- Add support for enterprise managed MCP allowlist and denylist
27
28## 2.0.21
29
30- Support MCP `structuredContent` field in tool responses
31- Added an interactive question tool
32- Claude will now ask you questions more often in plan mode
33- Added Haiku 4.5 as a model option for Pro users
34- Fixed an issue where queued commands don't have access to previous messages' output
35
36## 2.0.20
37
38- Added support for Claude Skills
39
40## 2.0.19
41
42- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
43- Fixed a bug where Haiku was unnecessarily called in print mode
44
45## 2.0.17
46
47- Added Haiku 4.5 to model selector!
48- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
49- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
50- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
51- OTEL: support HTTP_PROXY and HTTPS_PROXY
52- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
53
54## 2.0.15
55
56- Fixed bug with resuming where previously created files needed to be read again before writing
57- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
58
59## 2.0.14
60
61- Fix @-mentioning MCP servers to toggle them on/off
62- Improve permission checks for bash with inline env vars
63- Fix ultrathink + thinking toggle
64- Reduce unnecessary logins
65- Document --system-prompt
66- Several improvements to rendering
67- Plugins UI polish
68
69## 2.0.13
70
71- Fixed `/plugin` not working on native build
72
73## 2.0.12
74
75- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
76- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
77- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
78- `/plugin validate` command for validating plugin structure and configuration
79- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
80- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
81- Comprehensive error messages and diagnostics via `/doctor` command
82- Avoid flickering in `/model` selector
83- Improvements to `/help`
84- Avoid mentioning hooks in `/resume` summaries
85- Changes to the "verbose" setting in `/config` now persist across sessions
86
87## 2.0.11
88
89- Reduced system prompt size by 1.4k tokens
90- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
91- Fixed Opus fallback rate limit errors appearing incorrectly
92- Fixed /add-dir command selecting wrong default tab
93
94## 2.0.10
95
96- Rewrote terminal renderer for buttery smooth UI
97- Enable/disable MCP servers by @mentioning, or in /mcp
98- Added tab completion for shell commands in bash mode
99- PreToolUse hooks can now modify tool inputs
100- Press Ctrl-G to edit your prompt in your system's configured text editor
101- Fixes for bash permission checks with environment variables in the command
102
103## 2.0.9
104
105- Fix regression where bash backgrounding stopped working
106
107## 2.0.8
108
109- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
110- IDE: Add drag-and-drop support for files and folders in chat
111- /context: Fix counting for thinking blocks
112- Improve message rendering for users with light themes on dark terminals
113- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
114
115## 2.0.5
116
117- IDE: Fix IME unintended message submission with Enter and Tab
118- IDE: Add "Open in Terminal" link in login screen
119- Fix unhandled OAuth expiration 401 API errors
120- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
121
122## 2.0.1
123
124- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
125- Various bug fixes and presentation improvements
126
127## 2.0.0
128
129- New native VS Code extension
130- Fresh coat of paint throughout the whole app
131- /rewind a conversation to undo code changes
132- /usage command to see plan limits
133- Tab to toggle thinking (sticky across sessions)
134- Ctrl-R to search history
135- Unshipped claude config command
136- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
137- SDK: The Claude Code SDK is now the Claude Agent SDK
138- Add subagents dynamically with `--agents` flag
139
140## 1.0.126
141
142- Enable /context command for Bedrock and Vertex
143- Add mTLS support for HTTP-based OpenTelemetry exporters
144
145## 1.0.124
146
147- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
148- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
149- No longer inform Claude of the list of allowed tools when permission is denied
150- Fixed security vulnerability in Bash tool permission checks
151- Improved VSCode extension performance for large files
152
153## 1.0.123
154
155- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
156- Fixed thinking mode triggering on negation phrases like "don't think"
157- Fixed rendering performance degradation during token streaming
158- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
159- Enhanced BashTool environment snapshot logging
160- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
161- Migrated --debug logging to a file, to enable easy tailing & filtering
162
163## 1.0.120
164
165- Fix input lag during typing, especially noticeable with large prompts
166- Improved VSCode extension command registry and sessions dialog user experience
167- Enhanced sessions dialog responsiveness and visual feedback
168- Fixed IDE compatibility issue by removing worktree support check
169- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
170
171## 1.0.119
172
173- Fix Windows issue where process visually freezes on entering interactive mode
174- Support dynamic headers for MCP servers via headersHelper configuration
175- Fix thinking mode not working in headless sessions
176- Fix slash commands now properly update allowed tools instead of replacing them
177
178## 1.0.117
179
180- Add Ctrl-R history search to recall previous commands like bash/zsh
181- Fix input lag while typing, especially on Windows
182- Add sed command to auto-allowed commands in acceptEdits mode
183- Fix Windows PATH comparison to be case-insensitive for drive letters
184- Add permissions management hint to /add-dir output
185
186## 1.0.115
187
188- Improve thinking mode display with enhanced visual effects
189- Type /t to temporarily disable thinking mode in your prompt
190- Improve path validation for glob and grep tools
191- Show condensed output for post-tool hooks to reduce visual clutter
192- Fix visual feedback when loading state completes
193- Improve UI consistency for permission request dialogs
194
195## 1.0.113
196
197- Deprecated piped input in interactive mode
198- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
199
200## 1.0.112
201
202- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
203- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
204- Hooks: Added systemMessage support for SessionEnd hooks
205- Added `spinnerTipsEnabled` setting to disable spinner tips
206- IDE: Various improvements and bug fixes
207
208## 1.0.111
209
210- /model now validates provided model names
211- Fixed Bash tool crashes caused by malformed shell syntax parsing
212
213## 1.0.110
214
215- /terminal-setup command now supports WezTerm
216- MCP: OAuth tokens now proactively refresh before expiration
217- Fixed reliability issues with background Bash processes
218
219## 1.0.109
220
221- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
222
223## 1.0.106
224
225- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
226
227## 1.0.97
228
229- Settings: /doctor now validates permission rule syntax and suggests corrections
230
231## 1.0.94
232
233- Vertex: add support for global endpoints for supported models
234- /memory command now allows direct editing of all imported memory files
235- SDK: Add custom tools as callbacks
236- Added /todos command to list current todo items
237
238## 1.0.93
239
240- Windows: Add alt + v shortcut for pasting images from clipboard
241- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
242
243## 1.0.90
244
245- Settings file changes take effect immediately - no restart required
246
247## 1.0.88
248
249- Fixed issue causing "OAuth authentication is currently not supported"
250- Status line input now includes `exceeds_200k_tokens`
251- Fixed incorrect usage tracking in /cost.
252- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
253- Bedrock: Updated default Sonnet model to Sonnet 4
254
255## 1.0.86
256
257- Added /context to help users self-serve debug context issues
258- SDK: Added UUID support for all SDK messages
259- SDK: Added `--replay-user-messages` to replay user messages back to stdout
260
261## 1.0.85
262
263- Status line input now includes session cost info
264- Hooks: Introduced SessionEnd hook
265
266## 1.0.84
267
268- Fix tool_use/tool_result id mismatch error when network is unstable
269- Fix Claude sometimes ignoring real-time steering when wrapping up a task
270- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
271- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
272
273## 1.0.83
274
275- @-mention: Support files with spaces in path
276- New shimmering spinner
277
278## 1.0.82
279
280- SDK: Add request cancellation support
281- SDK: New additionalDirectories option to search custom paths, improved slash command processing
282- Settings: Validation prevents invalid fields in .claude/settings.json files
283- MCP: Improve tool name consistency
284- Bash: Fix crash when Claude tries to automatically read large files
285
286## 1.0.81
287
288- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
289- Agents: Fix custom agent loading when agent files are unparsable
290
291## 1.0.80
292
293- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
294
295## 1.0.77
296
297- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
298- SDK: Add session support and permission denial tracking
299- Fix token limit errors in conversation summarization
300- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
301
302## 1.0.73
303
304- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
305- MCP: Press Esc to cancel OAuth authentication flows
306- Bash: Improved command validation and reduced false security warnings
307- UI: Enhanced spinner animations and status line visual hierarchy
308- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
309
310## 1.0.72
311
312- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
313
314## 1.0.71
315
316- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
317- Customizable status line: add your terminal prompt to Claude Code with /statusline
318
319## 1.0.70
320
321- Performance: Optimized message rendering for better performance with large contexts
322- Windows: Fixed native file search, ripgrep, and subagent functionality
323- Added support for @-mentions in slash command arguments
324
325## 1.0.69
326
327- Upgraded Opus to version 4.1
328
329## 1.0.68
330
331- Fix incorrect model names being used for certain commands like `/pr-comments`
332- 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.
333- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
334- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
335- SDK: Added canUseTool callback support for tool confirmation
336- Added `disableAllHooks` setting
337- Improved file suggestions performance in large repos
338
339## 1.0.65
340
341- IDE: Fixed connection stability issues and error handling for diagnostics
342- Windows: Fixed shell environment setup for users without .bashrc files
343
344## 1.0.64
345
346- Agents: Added model customization support - you can now specify which model an agent should use
347- Agents: Fixed unintended access to the recursive agent tool
348- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
349- SDK: Fixed user input tracking across multi-turn conversations
350- Added hidden files to file search and @-mention suggestions
351
352## 1.0.63
353
354- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
355
356## 1.0.62
357
358- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
359- Hooks: Added SessionStart hook for new session initialization
360- /add-dir command now supports typeahead for directory paths
361- Improved network connectivity check reliability
362
363## 1.0.61
364
365- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
366- Settings: Added `--settings` flag to load settings from a JSON file
367- Settings: Fixed resolution of settings files paths that are symlinks
368- OTEL: Fixed reporting of wrong organization after authentication changes
369- Slash commands: Fixed permissions checking for allowed-tools with Bash
370- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
371- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
372- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
373
374## 1.0.60
375
376- You can now create custom subagents for specialized tasks! Run /agents to get started
377
378## 1.0.59
379
380- SDK: Added tool confirmation support with canUseTool callback
381- SDK: Allow specifying env for spawned process
382- Hooks: Exposed PermissionDecision to hooks (including "ask")
383- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
384- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
385
386## 1.0.58
387
388- Added support for reading PDFs
389- MCP: Improved server health status display in 'claude mcp list'
390- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
391
392## 1.0.57
393
394- Added support for specifying a model in slash commands
395- Improved permission messages to help Claude understand allowed tools
396- Fix: Remove trailing newlines from bash output in terminal wrapping
397
398## 1.0.56
399
400- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
401- Fixes for WSL IDE detection
402- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
403
404## 1.0.55
405
406- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
407- Windows: fixed Ctrl+Z crash
408- SDK: Added ability to capture error logging
409- Add --system-prompt-file option to override system prompt in print mode
410
411## 1.0.54
412
413- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
414- Custom slash commands: Added argument-hint to frontmatter
415- Windows: OAuth uses port 45454 and properly constructs browser URL
416- Windows: mode switching now uses alt + m, and plan mode renders properly
417- Shell: Switch to in-memory shell snapshot to fix file-related errors
418
419## 1.0.53
420
421- Updated @-mention file truncation from 100 lines to 2000 lines
422- 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).
423
424## 1.0.52
425
426- Added support for MCP server instructions
427
428## 1.0.51
429
430- Added support for native Windows (requires Git for Windows)
431- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
432- Settings: /doctor can now help you identify and fix invalid setting files
433- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
434- Increased auto-compact warning threshold from 60% to 80%
435- Fixed an issue with handling user directories with spaces for shell snapshots
436- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
437- Custom slash commands: Fixed user-level commands in subdirectories
438- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
439
440## 1.0.48
441
442- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
443- Added progress messages to Bash tool based on the last 5 lines of command output
444- Added expanding variables support for MCP server configuration
445- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
446- Improved IDE extension path handling when Claude Code runs in WSL
447- Hooks: Added a PreCompact hook
448- Vim mode: Added c, f/F, t/T
449
450## 1.0.45
451
452- Redesigned Search (Grep) tool with new tool input parameters and features
453- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
454- Fixed config file corruption issue by enforcing atomic writes
455- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
456- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
457- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
458
459## 1.0.44
460
461- New /export command lets you quickly export a conversation for sharing
462- MCP: resource_link tool results are now supported
463- MCP: tool annotations and tool titles now display in /mcp view
464- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
465
466## 1.0.43
467
468- Fixed a bug where the theme selector was saving excessively
469- Hooks: Added EPIPE system error handling
470
471## 1.0.42
472
473- Added tilde (`~`) expansion support to `/add-dir` command
474
475## 1.0.41
476
477- Hooks: Split Stop hook triggering into Stop and SubagentStop
478- Hooks: Enabled optional timeout configuration for each command
479- Hooks: Added "hook_event_name" to hook input
480- Fixed a bug where MCP tools would display twice in tool list
481- New tool parameters JSON for Bash tool in `tool_decision` event
482
483## 1.0.40
484
485- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
486
487## 1.0.39
488
489- New Active Time metric in OpenTelemetry logging
490
491## 1.0.38
492
493- 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
494
495## 1.0.37
496
497- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
498
499## 1.0.36
500
501- Web search now takes today's date into context
502- Fixed a bug where stdio MCP servers were not terminating properly on exit
503
504## 1.0.35
505
506- Added support for MCP OAuth Authorization Server discovery
507
508## 1.0.34
509
510- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
511
512## 1.0.33
513
514- Improved logging functionality with session ID support
515- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
516- Improvements to plan mode
517
518## 1.0.32
519
520- Updated loopback config for litellm
521- Added forceLoginMethod setting to bypass login selection screen
522
523## 1.0.31
524
525- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
526
527## 1.0.30
528
529- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
530- Improved file path autocomplete with filename matching
531- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
532- Enhanced jq regex support for complex filters with pipes and select
533
534## 1.0.29
535
536- Improved CJK character support in cursor navigation and rendering
537
538## 1.0.28
539
540- Slash commands: Fix selector display during history navigation
541- Resizes images before upload to prevent API size limit errors
542- Added XDG_CONFIG_HOME support to configuration directory
543- Performance optimizations for memory usage
544- New attributes (terminal.type, language) in OpenTelemetry logging
545
546## 1.0.27
547
548- Streamable HTTP MCP servers are now supported
549- Remote MCP servers (SSE and HTTP) now support OAuth
550- MCP resources can now be @-mentioned
551- /resume slash command to switch conversations within Claude Code
552
553## 1.0.25
554
555- Slash commands: moved "project" and "user" prefixes to descriptions
556- Slash commands: improved reliability for command discovery
557- Improved support for Ghostty
558- Improved web search reliability
559
560## 1.0.24
561
562- Improved /mcp output
563- Fixed a bug where settings arrays got overwritten instead of merged
564
565## 1.0.23
566
567- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
568- Released Python SDK: pip install claude-code-sdk to get started
569
570## 1.0.22
571
572- SDK: Renamed `total_cost` to `total_cost_usd`
573
574## 1.0.21
575
576- Improved editing of files with tab-based indentation
577- Fix for tool_use without matching tool_result errors
578- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
579
580## 1.0.18
581
582- Added --add-dir CLI argument for specifying additional working directories
583- Added streaming input support without require -p flag
584- Improved startup performance and session storage performance
585- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
586- Added detailed MCP server tools display (/mcp)
587- MCP authentication and permission improvements
588- Added auto-reconnection for MCP SSE connections on disconnect
589- Fixed issue where pasted content was lost when dialogs appeared
590
591## 1.0.17
592
593- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
594- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
595- MCP server list UI improvements
596- Update Claude Code process title to display "claude" instead of "node"
597
598## 1.0.11
599
600- Claude Code can now also be used with a Claude Pro subscription
601- Added /upgrade for smoother switching to Claude Max plans
602- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
603- Improved shell configuration error handling
604- Improved todo list handling during compaction
605
606## 1.0.10
607
608- Added markdown table support
609- Improved streaming performance
610
611## 1.0.8
612
613- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
614- Increased default otel interval from 1s -> 5s
615- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
616- Fixed a regression where search tools unnecessarily asked for permissions
617- Added support for triggering thinking non-English languages
618- Improved compacting UI
619
620## 1.0.7
621
622- Renamed /allowed-tools -> /permissions
623- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
624- Deprecated claude config commands in favor of editing settings.json
625- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
626- Improved error handling for /install-github-app
627- Bugfixes, UI polish, and tool reliability improvements
628
629## 1.0.6
630
631- Improved edit reliability for tab-indented files
632- Respect CLAUDE_CONFIG_DIR everywhere
633- Reduced unnecessary tool permission prompts
634- Added support for symlinks in @file typeahead
635- Bugfixes, UI polish, and tool reliability improvements
636
637## 1.0.4
638
639- Fixed a bug where MCP tool errors weren't being parsed correctly
640
641## 1.0.1
642
643- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
644- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
645- Updated documentation links and OAuth process descriptions
646
647## 1.0.0
648
649- Claude Code is now generally available
650- Introducing Sonnet 4 and Opus 4 models
651
652## 0.2.125
653
654- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
655- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
656
657## 0.2.117
658
659- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
660- Introduced settings.cleanupPeriodDays
661- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
662- Introduced --debug mode
663
664## 0.2.108
665
666- You can now send messages to Claude while it works to steer Claude in real-time
667- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
668- Fixed a bug where thinking was not working in -p mode
669- Fixed a regression in /cost reporting
670- Deprecated MCP wizard interface in favor of other MCP commands
671- Lots of other bugfixes and improvements
672
673## 0.2.107
674
675- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
676
677## 0.2.106
678
679- MCP SSE server configs can now specify custom headers
680- Fixed a bug where MCP permission prompt didn't always show correctly
681
682## 0.2.105
683
684- Claude can now search the web
685- Moved system & account status to /status
686- Added word movement keybindings for Vim
687- Improved latency for startup, todo tool, and file edits
688
689## 0.2.102
690
691- Improved thinking triggering reliability
692- Improved @mention reliability for images and folders
693- You can now paste multiple large chunks into one prompt
694
695## 0.2.100
696
697- Fixed a crash caused by a stack overflow error
698- Made db storage optional; missing db support disables --continue and --resume
699
700## 0.2.98
701
702- Fixed an issue where auto-compact was running twice
703
704## 0.2.96
705
706- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
707
708## 0.2.93
709
710- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
711- Claude now has access to a Todo list that helps it stay on track and be more organized
712
713## 0.2.82
714
715- Added support for --disallowedTools
716- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
717
718## 0.2.75
719
720- Hit Enter to queue up additional messages while Claude is working
721- Drag in or copy/paste image files directly into the prompt
722- @-mention files to directly add them to context
723- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
724- Improved performance for filename auto-complete
725
726## 0.2.74
727
728- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
729- Task tool can now perform writes and run bash commands
730
731## 0.2.72
732
733- Updated spinner to indicate tokens loaded and tool usage
734
735## 0.2.70
736
737- Network commands like curl are now available for Claude to use
738- Claude can now run multiple web queries in parallel
739- Pressing ESC once immediately interrupts Claude in Auto-accept mode
740
741## 0.2.69
742
743- Fixed UI glitches with improved Select component behavior
744- Enhanced terminal output display with better text truncation logic
745
746## 0.2.67
747
748- Shared project permission rules can be saved in .claude/settings.json
749
750## 0.2.66
751
752- Print mode (-p) now supports streaming output via --output-format=stream-json
753- Fixed issue where pasting could trigger memory or bash mode unexpectedly
754
755## 0.2.63
756
757- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
758
759## 0.2.61
760
761- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
762- Enhanced image detection for more reliable clipboard paste functionality
763- Fixed an issue where ESC key could crash the conversation history selector
764
765## 0.2.59
766
767- Copy+paste images directly into your prompt
768- Improved progress indicators for bash and fetch tools
769- Bugfixes for non-interactive mode (-p)
770
771## 0.2.54
772
773- Quickly add to Memory by starting your message with '#'
774- Press ctrl+r to see full output for long tool results
775- Added support for MCP SSE transport
776
777## 0.2.53
778
779- New web fetch tool lets Claude view URLs that you paste in
780- Fixed a bug with JPEG detection
781
782## 0.2.50
783
784- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
785
786## 0.2.49
787
788- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
789
790## 0.2.47
791
792- Press Tab to auto-complete file and folder names
793- Press Shift + Tab to toggle auto-accept for file edits
794- Automatic conversation compaction for infinite conversation length (toggle with /config)
795
796## 0.2.44
797
798- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
799
800## 0.2.41
801
802- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
803- MCP server startup no longer blocks the app from starting up
804
805## 0.2.37
806
807- New /release-notes command lets you view release notes at any time
808- `claude config add/remove` commands now accept multiple values separated by commas or spaces
809
810## 0.2.36
811
812- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
813- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
814
815## 0.2.34
816
817- Vim bindings for text input - enable with /vim or /config
818
819## 0.2.32
820
821- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
822- Fix for some PersistentShell issues
823
824## 0.2.31
825
826- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
827- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
828
829## 0.2.30
830
831- Added ANSI color theme for better terminal compatibility
832- Fixed issue where slash command arguments weren't being sent properly
833- (Mac-only) API keys are now stored in macOS Keychain
834
835## 0.2.26
836
837- New /approved-tools command for managing tool permissions
838- Word-level diff display for improved code readability
839- Fuzzy matching for slash commands
840
841## 0.2.21
842
843- Fuzzy matching for /commands