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 3be7215 · on Oct 21, 2025 History
835 lines · 30.0 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.25
4
5- 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
6
7## 2.0.24
8
9- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
10- Claude Code Web: Support for Web -> CLI teleport
11- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
12
13## 2.0.22
14
15- Fixed content layout shift when scrolling through slash commands
16- IDE: Add toggle to enable/disable thinking.
17- Fix bug causing duplicate permission prompts with parallel tool calls
18- Add support for enterprise managed MCP allowlist and denylist
19
20## 2.0.21
21
22- Support MCP `structuredContent` field in tool responses
23- Added an interactive question tool
24- Claude will now ask you questions more often in plan mode
25- Added Haiku 4.5 as a model option for Pro users
26- Fixed an issue where queued commands don't have access to previous messages' output
27
28## 2.0.20
29
30- Added support for Claude Skills
31
32## 2.0.19
33
34- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
35- Fixed a bug where Haiku was unnecessarily called in print mode
36
37## 2.0.17
38
39- Added Haiku 4.5 to model selector!
40- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
41- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
42- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
43- OTEL: support HTTP_PROXY and HTTPS_PROXY
44- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
45
46## 2.0.15
47
48- Fixed bug with resuming where previously created files needed to be read again before writing
49- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
50
51## 2.0.14
52
53- Fix @-mentioning MCP servers to toggle them on/off
54- Improve permission checks for bash with inline env vars
55- Fix ultrathink + thinking toggle
56- Reduce unnecessary logins
57- Document --system-prompt
58- Several improvements to rendering
59- Plugins UI polish
60
61## 2.0.13
62
63- Fixed `/plugin` not working on native build
64
65## 2.0.12
66
67- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
68- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
69- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
70- `/plugin validate` command for validating plugin structure and configuration
71- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
72- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
73- Comprehensive error messages and diagnostics via `/doctor` command
74- Avoid flickering in `/model` selector
75- Improvements to `/help`
76- Avoid mentioning hooks in `/resume` summaries
77- Changes to the "verbose" setting in `/config` now persist across sessions
78
79## 2.0.11
80
81- Reduced system prompt size by 1.4k tokens
82- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
83- Fixed Opus fallback rate limit errors appearing incorrectly
84- Fixed /add-dir command selecting wrong default tab
85
86## 2.0.10
87
88- Rewrote terminal renderer for buttery smooth UI
89- Enable/disable MCP servers by @mentioning, or in /mcp
90- Added tab completion for shell commands in bash mode
91- PreToolUse hooks can now modify tool inputs
92- Press Ctrl-G to edit your prompt in your system's configured text editor
93- Fixes for bash permission checks with environment variables in the command
94
95## 2.0.9
96
97- Fix regression where bash backgrounding stopped working
98
99## 2.0.8
100
101- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
102- IDE: Add drag-and-drop support for files and folders in chat
103- /context: Fix counting for thinking blocks
104- Improve message rendering for users with light themes on dark terminals
105- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
106
107## 2.0.5
108
109- IDE: Fix IME unintended message submission with Enter and Tab
110- IDE: Add "Open in Terminal" link in login screen
111- Fix unhandled OAuth expiration 401 API errors
112- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
113
114## 2.0.1
115
116- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
117- Various bug fixes and presentation improvements
118
119## 2.0.0
120
121- New native VS Code extension
122- Fresh coat of paint throughout the whole app
123- /rewind a conversation to undo code changes
124- /usage command to see plan limits
125- Tab to toggle thinking (sticky across sessions)
126- Ctrl-R to search history
127- Unshipped claude config command
128- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
129- SDK: The Claude Code SDK is now the Claude Agent SDK
130- Add subagents dynamically with `--agents` flag
131
132## 1.0.126
133
134- Enable /context command for Bedrock and Vertex
135- Add mTLS support for HTTP-based OpenTelemetry exporters
136
137## 1.0.124
138
139- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
140- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
141- No longer inform Claude of the list of allowed tools when permission is denied
142- Fixed security vulnerability in Bash tool permission checks
143- Improved VSCode extension performance for large files
144
145## 1.0.123
146
147- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
148- Fixed thinking mode triggering on negation phrases like "don't think"
149- Fixed rendering performance degradation during token streaming
150- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
151- Enhanced BashTool environment snapshot logging
152- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
153- Migrated --debug logging to a file, to enable easy tailing & filtering
154
155## 1.0.120
156
157- Fix input lag during typing, especially noticeable with large prompts
158- Improved VSCode extension command registry and sessions dialog user experience
159- Enhanced sessions dialog responsiveness and visual feedback
160- Fixed IDE compatibility issue by removing worktree support check
161- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
162
163## 1.0.119
164
165- Fix Windows issue where process visually freezes on entering interactive mode
166- Support dynamic headers for MCP servers via headersHelper configuration
167- Fix thinking mode not working in headless sessions
168- Fix slash commands now properly update allowed tools instead of replacing them
169
170## 1.0.117
171
172- Add Ctrl-R history search to recall previous commands like bash/zsh
173- Fix input lag while typing, especially on Windows
174- Add sed command to auto-allowed commands in acceptEdits mode
175- Fix Windows PATH comparison to be case-insensitive for drive letters
176- Add permissions management hint to /add-dir output
177
178## 1.0.115
179
180- Improve thinking mode display with enhanced visual effects
181- Type /t to temporarily disable thinking mode in your prompt
182- Improve path validation for glob and grep tools
183- Show condensed output for post-tool hooks to reduce visual clutter
184- Fix visual feedback when loading state completes
185- Improve UI consistency for permission request dialogs
186
187## 1.0.113
188
189- Deprecated piped input in interactive mode
190- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
191
192## 1.0.112
193
194- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
195- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
196- Hooks: Added systemMessage support for SessionEnd hooks
197- Added `spinnerTipsEnabled` setting to disable spinner tips
198- IDE: Various improvements and bug fixes
199
200## 1.0.111
201
202- /model now validates provided model names
203- Fixed Bash tool crashes caused by malformed shell syntax parsing
204
205## 1.0.110
206
207- /terminal-setup command now supports WezTerm
208- MCP: OAuth tokens now proactively refresh before expiration
209- Fixed reliability issues with background Bash processes
210
211## 1.0.109
212
213- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
214
215## 1.0.106
216
217- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
218
219## 1.0.97
220
221- Settings: /doctor now validates permission rule syntax and suggests corrections
222
223## 1.0.94
224
225- Vertex: add support for global endpoints for supported models
226- /memory command now allows direct editing of all imported memory files
227- SDK: Add custom tools as callbacks
228- Added /todos command to list current todo items
229
230## 1.0.93
231
232- Windows: Add alt + v shortcut for pasting images from clipboard
233- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
234
235## 1.0.90
236
237- Settings file changes take effect immediately - no restart required
238
239## 1.0.88
240
241- Fixed issue causing "OAuth authentication is currently not supported"
242- Status line input now includes `exceeds_200k_tokens`
243- Fixed incorrect usage tracking in /cost.
244- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
245- Bedrock: Updated default Sonnet model to Sonnet 4
246
247## 1.0.86
248
249- Added /context to help users self-serve debug context issues
250- SDK: Added UUID support for all SDK messages
251- SDK: Added `--replay-user-messages` to replay user messages back to stdout
252
253## 1.0.85
254
255- Status line input now includes session cost info
256- Hooks: Introduced SessionEnd hook
257
258## 1.0.84
259
260- Fix tool_use/tool_result id mismatch error when network is unstable
261- Fix Claude sometimes ignoring real-time steering when wrapping up a task
262- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
263- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
264
265## 1.0.83
266
267- @-mention: Support files with spaces in path
268- New shimmering spinner
269
270## 1.0.82
271
272- SDK: Add request cancellation support
273- SDK: New additionalDirectories option to search custom paths, improved slash command processing
274- Settings: Validation prevents invalid fields in .claude/settings.json files
275- MCP: Improve tool name consistency
276- Bash: Fix crash when Claude tries to automatically read large files
277
278## 1.0.81
279
280- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
281- Agents: Fix custom agent loading when agent files are unparsable
282
283## 1.0.80
284
285- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
286
287## 1.0.77
288
289- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
290- SDK: Add session support and permission denial tracking
291- Fix token limit errors in conversation summarization
292- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
293
294## 1.0.73
295
296- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
297- MCP: Press Esc to cancel OAuth authentication flows
298- Bash: Improved command validation and reduced false security warnings
299- UI: Enhanced spinner animations and status line visual hierarchy
300- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
301
302## 1.0.72
303
304- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
305
306## 1.0.71
307
308- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
309- Customizable status line: add your terminal prompt to Claude Code with /statusline
310
311## 1.0.70
312
313- Performance: Optimized message rendering for better performance with large contexts
314- Windows: Fixed native file search, ripgrep, and subagent functionality
315- Added support for @-mentions in slash command arguments
316
317## 1.0.69
318
319- Upgraded Opus to version 4.1
320
321## 1.0.68
322
323- Fix incorrect model names being used for certain commands like `/pr-comments`
324- 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.
325- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
326- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
327- SDK: Added canUseTool callback support for tool confirmation
328- Added `disableAllHooks` setting
329- Improved file suggestions performance in large repos
330
331## 1.0.65
332
333- IDE: Fixed connection stability issues and error handling for diagnostics
334- Windows: Fixed shell environment setup for users without .bashrc files
335
336## 1.0.64
337
338- Agents: Added model customization support - you can now specify which model an agent should use
339- Agents: Fixed unintended access to the recursive agent tool
340- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
341- SDK: Fixed user input tracking across multi-turn conversations
342- Added hidden files to file search and @-mention suggestions
343
344## 1.0.63
345
346- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
347
348## 1.0.62
349
350- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
351- Hooks: Added SessionStart hook for new session initialization
352- /add-dir command now supports typeahead for directory paths
353- Improved network connectivity check reliability
354
355## 1.0.61
356
357- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
358- Settings: Added `--settings` flag to load settings from a JSON file
359- Settings: Fixed resolution of settings files paths that are symlinks
360- OTEL: Fixed reporting of wrong organization after authentication changes
361- Slash commands: Fixed permissions checking for allowed-tools with Bash
362- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
363- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
364- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
365
366## 1.0.60
367
368- You can now create custom subagents for specialized tasks! Run /agents to get started
369
370## 1.0.59
371
372- SDK: Added tool confirmation support with canUseTool callback
373- SDK: Allow specifying env for spawned process
374- Hooks: Exposed PermissionDecision to hooks (including "ask")
375- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
376- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
377
378## 1.0.58
379
380- Added support for reading PDFs
381- MCP: Improved server health status display in 'claude mcp list'
382- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
383
384## 1.0.57
385
386- Added support for specifying a model in slash commands
387- Improved permission messages to help Claude understand allowed tools
388- Fix: Remove trailing newlines from bash output in terminal wrapping
389
390## 1.0.56
391
392- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
393- Fixes for WSL IDE detection
394- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
395
396## 1.0.55
397
398- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
399- Windows: fixed Ctrl+Z crash
400- SDK: Added ability to capture error logging
401- Add --system-prompt-file option to override system prompt in print mode
402
403## 1.0.54
404
405- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
406- Custom slash commands: Added argument-hint to frontmatter
407- Windows: OAuth uses port 45454 and properly constructs browser URL
408- Windows: mode switching now uses alt + m, and plan mode renders properly
409- Shell: Switch to in-memory shell snapshot to fix file-related errors
410
411## 1.0.53
412
413- Updated @-mention file truncation from 100 lines to 2000 lines
414- 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).
415
416## 1.0.52
417
418- Added support for MCP server instructions
419
420## 1.0.51
421
422- Added support for native Windows (requires Git for Windows)
423- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
424- Settings: /doctor can now help you identify and fix invalid setting files
425- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
426- Increased auto-compact warning threshold from 60% to 80%
427- Fixed an issue with handling user directories with spaces for shell snapshots
428- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
429- Custom slash commands: Fixed user-level commands in subdirectories
430- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
431
432## 1.0.48
433
434- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
435- Added progress messages to Bash tool based on the last 5 lines of command output
436- Added expanding variables support for MCP server configuration
437- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
438- Improved IDE extension path handling when Claude Code runs in WSL
439- Hooks: Added a PreCompact hook
440- Vim mode: Added c, f/F, t/T
441
442## 1.0.45
443
444- Redesigned Search (Grep) tool with new tool input parameters and features
445- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
446- Fixed config file corruption issue by enforcing atomic writes
447- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
448- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
449- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
450
451## 1.0.44
452
453- New /export command lets you quickly export a conversation for sharing
454- MCP: resource_link tool results are now supported
455- MCP: tool annotations and tool titles now display in /mcp view
456- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
457
458## 1.0.43
459
460- Fixed a bug where the theme selector was saving excessively
461- Hooks: Added EPIPE system error handling
462
463## 1.0.42
464
465- Added tilde (`~`) expansion support to `/add-dir` command
466
467## 1.0.41
468
469- Hooks: Split Stop hook triggering into Stop and SubagentStop
470- Hooks: Enabled optional timeout configuration for each command
471- Hooks: Added "hook_event_name" to hook input
472- Fixed a bug where MCP tools would display twice in tool list
473- New tool parameters JSON for Bash tool in `tool_decision` event
474
475## 1.0.40
476
477- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
478
479## 1.0.39
480
481- New Active Time metric in OpenTelemetry logging
482
483## 1.0.38
484
485- 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
486
487## 1.0.37
488
489- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
490
491## 1.0.36
492
493- Web search now takes today's date into context
494- Fixed a bug where stdio MCP servers were not terminating properly on exit
495
496## 1.0.35
497
498- Added support for MCP OAuth Authorization Server discovery
499
500## 1.0.34
501
502- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
503
504## 1.0.33
505
506- Improved logging functionality with session ID support
507- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
508- Improvements to plan mode
509
510## 1.0.32
511
512- Updated loopback config for litellm
513- Added forceLoginMethod setting to bypass login selection screen
514
515## 1.0.31
516
517- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
518
519## 1.0.30
520
521- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
522- Improved file path autocomplete with filename matching
523- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
524- Enhanced jq regex support for complex filters with pipes and select
525
526## 1.0.29
527
528- Improved CJK character support in cursor navigation and rendering
529
530## 1.0.28
531
532- Slash commands: Fix selector display during history navigation
533- Resizes images before upload to prevent API size limit errors
534- Added XDG_CONFIG_HOME support to configuration directory
535- Performance optimizations for memory usage
536- New attributes (terminal.type, language) in OpenTelemetry logging
537
538## 1.0.27
539
540- Streamable HTTP MCP servers are now supported
541- Remote MCP servers (SSE and HTTP) now support OAuth
542- MCP resources can now be @-mentioned
543- /resume slash command to switch conversations within Claude Code
544
545## 1.0.25
546
547- Slash commands: moved "project" and "user" prefixes to descriptions
548- Slash commands: improved reliability for command discovery
549- Improved support for Ghostty
550- Improved web search reliability
551
552## 1.0.24
553
554- Improved /mcp output
555- Fixed a bug where settings arrays got overwritten instead of merged
556
557## 1.0.23
558
559- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
560- Released Python SDK: pip install claude-code-sdk to get started
561
562## 1.0.22
563
564- SDK: Renamed `total_cost` to `total_cost_usd`
565
566## 1.0.21
567
568- Improved editing of files with tab-based indentation
569- Fix for tool_use without matching tool_result errors
570- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
571
572## 1.0.18
573
574- Added --add-dir CLI argument for specifying additional working directories
575- Added streaming input support without require -p flag
576- Improved startup performance and session storage performance
577- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
578- Added detailed MCP server tools display (/mcp)
579- MCP authentication and permission improvements
580- Added auto-reconnection for MCP SSE connections on disconnect
581- Fixed issue where pasted content was lost when dialogs appeared
582
583## 1.0.17
584
585- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
586- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
587- MCP server list UI improvements
588- Update Claude Code process title to display "claude" instead of "node"
589
590## 1.0.11
591
592- Claude Code can now also be used with a Claude Pro subscription
593- Added /upgrade for smoother switching to Claude Max plans
594- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
595- Improved shell configuration error handling
596- Improved todo list handling during compaction
597
598## 1.0.10
599
600- Added markdown table support
601- Improved streaming performance
602
603## 1.0.8
604
605- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
606- Increased default otel interval from 1s -> 5s
607- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
608- Fixed a regression where search tools unnecessarily asked for permissions
609- Added support for triggering thinking non-English languages
610- Improved compacting UI
611
612## 1.0.7
613
614- Renamed /allowed-tools -> /permissions
615- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
616- Deprecated claude config commands in favor of editing settings.json
617- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
618- Improved error handling for /install-github-app
619- Bugfixes, UI polish, and tool reliability improvements
620
621## 1.0.6
622
623- Improved edit reliability for tab-indented files
624- Respect CLAUDE_CONFIG_DIR everywhere
625- Reduced unnecessary tool permission prompts
626- Added support for symlinks in @file typeahead
627- Bugfixes, UI polish, and tool reliability improvements
628
629## 1.0.4
630
631- Fixed a bug where MCP tool errors weren't being parsed correctly
632
633## 1.0.1
634
635- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
636- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
637- Updated documentation links and OAuth process descriptions
638
639## 1.0.0
640
641- Claude Code is now generally available
642- Introducing Sonnet 4 and Opus 4 models
643
644## 0.2.125
645
646- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
647- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
648
649## 0.2.117
650
651- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
652- Introduced settings.cleanupPeriodDays
653- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
654- Introduced --debug mode
655
656## 0.2.108
657
658- You can now send messages to Claude while it works to steer Claude in real-time
659- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
660- Fixed a bug where thinking was not working in -p mode
661- Fixed a regression in /cost reporting
662- Deprecated MCP wizard interface in favor of other MCP commands
663- Lots of other bugfixes and improvements
664
665## 0.2.107
666
667- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
668
669## 0.2.106
670
671- MCP SSE server configs can now specify custom headers
672- Fixed a bug where MCP permission prompt didn't always show correctly
673
674## 0.2.105
675
676- Claude can now search the web
677- Moved system & account status to /status
678- Added word movement keybindings for Vim
679- Improved latency for startup, todo tool, and file edits
680
681## 0.2.102
682
683- Improved thinking triggering reliability
684- Improved @mention reliability for images and folders
685- You can now paste multiple large chunks into one prompt
686
687## 0.2.100
688
689- Fixed a crash caused by a stack overflow error
690- Made db storage optional; missing db support disables --continue and --resume
691
692## 0.2.98
693
694- Fixed an issue where auto-compact was running twice
695
696## 0.2.96
697
698- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
699
700## 0.2.93
701
702- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
703- Claude now has access to a Todo list that helps it stay on track and be more organized
704
705## 0.2.82
706
707- Added support for --disallowedTools
708- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
709
710## 0.2.75
711
712- Hit Enter to queue up additional messages while Claude is working
713- Drag in or copy/paste image files directly into the prompt
714- @-mention files to directly add them to context
715- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
716- Improved performance for filename auto-complete
717
718## 0.2.74
719
720- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
721- Task tool can now perform writes and run bash commands
722
723## 0.2.72
724
725- Updated spinner to indicate tokens loaded and tool usage
726
727## 0.2.70
728
729- Network commands like curl are now available for Claude to use
730- Claude can now run multiple web queries in parallel
731- Pressing ESC once immediately interrupts Claude in Auto-accept mode
732
733## 0.2.69
734
735- Fixed UI glitches with improved Select component behavior
736- Enhanced terminal output display with better text truncation logic
737
738## 0.2.67
739
740- Shared project permission rules can be saved in .claude/settings.json
741
742## 0.2.66
743
744- Print mode (-p) now supports streaming output via --output-format=stream-json
745- Fixed issue where pasting could trigger memory or bash mode unexpectedly
746
747## 0.2.63
748
749- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
750
751## 0.2.61
752
753- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
754- Enhanced image detection for more reliable clipboard paste functionality
755- Fixed an issue where ESC key could crash the conversation history selector
756
757## 0.2.59
758
759- Copy+paste images directly into your prompt
760- Improved progress indicators for bash and fetch tools
761- Bugfixes for non-interactive mode (-p)
762
763## 0.2.54
764
765- Quickly add to Memory by starting your message with '#'
766- Press ctrl+r to see full output for long tool results
767- Added support for MCP SSE transport
768
769## 0.2.53
770
771- New web fetch tool lets Claude view URLs that you paste in
772- Fixed a bug with JPEG detection
773
774## 0.2.50
775
776- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
777
778## 0.2.49
779
780- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
781
782## 0.2.47
783
784- Press Tab to auto-complete file and folder names
785- Press Shift + Tab to toggle auto-accept for file edits
786- Automatic conversation compaction for infinite conversation length (toggle with /config)
787
788## 0.2.44
789
790- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
791
792## 0.2.41
793
794- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
795- MCP server startup no longer blocks the app from starting up
796
797## 0.2.37
798
799- New /release-notes command lets you view release notes at any time
800- `claude config add/remove` commands now accept multiple values separated by commas or spaces
801
802## 0.2.36
803
804- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
805- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
806
807## 0.2.34
808
809- Vim bindings for text input - enable with /vim or /config
810
811## 0.2.32
812
813- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
814- Fix for some PersistentShell issues
815
816## 0.2.31
817
818- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
819- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
820
821## 0.2.30
822
823- Added ANSI color theme for better terminal compatibility
824- Fixed issue where slash command arguments weren't being sent properly
825- (Mac-only) API keys are now stored in macOS Keychain
826
827## 0.2.26
828
829- New /approved-tools command for managing tool permissions
830- Word-level diff display for improved code readability
831- Fuzzy matching for slash commands
832
833## 0.2.21
834
835- Fuzzy matching for /commands