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 ff0aafa · on Oct 20, 2025 History
831 lines · 29.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.24
4
5- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
6- Claude Code Web: Support for Web -> CLI teleport
7- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
8
9## 2.0.22
10
11- Fixed content layout shift when scrolling through slash commands
12- IDE: Add toggle to enable/disable thinking.
13- Fix bug causing duplicate permission prompts with parallel tool calls
14- Add support for enterprise managed MCP allowlist and denylist
15
16## 2.0.21
17
18- Support MCP `structuredContent` field in tool responses
19- Added an interactive question tool
20- Claude will now ask you questions more often in plan mode
21- Added Haiku 4.5 as a model option for Pro users
22- Fixed an issue where queued commands don't have access to previous messages' output
23
24## 2.0.20
25
26- Added support for Claude Skills
27
28## 2.0.19
29
30- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
31- Fixed a bug where Haiku was unnecessarily called in print mode
32
33## 2.0.17
34
35- Added Haiku 4.5 to model selector!
36- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
37- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
38- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
39- OTEL: support HTTP_PROXY and HTTPS_PROXY
40- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
41
42## 2.0.15
43
44- Fixed bug with resuming where previously created files needed to be read again before writing
45- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
46
47## 2.0.14
48
49- Fix @-mentioning MCP servers to toggle them on/off
50- Improve permission checks for bash with inline env vars
51- Fix ultrathink + thinking toggle
52- Reduce unnecessary logins
53- Document --system-prompt
54- Several improvements to rendering
55- Plugins UI polish
56
57## 2.0.13
58
59- Fixed `/plugin` not working on native build
60
61## 2.0.12
62
63- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
64- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
65- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
66- `/plugin validate` command for validating plugin structure and configuration
67- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
68- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
69- Comprehensive error messages and diagnostics via `/doctor` command
70- Avoid flickering in `/model` selector
71- Improvements to `/help`
72- Avoid mentioning hooks in `/resume` summaries
73- Changes to the "verbose" setting in `/config` now persist across sessions
74
75## 2.0.11
76
77- Reduced system prompt size by 1.4k tokens
78- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
79- Fixed Opus fallback rate limit errors appearing incorrectly
80- Fixed /add-dir command selecting wrong default tab
81
82## 2.0.10
83
84- Rewrote terminal renderer for buttery smooth UI
85- Enable/disable MCP servers by @mentioning, or in /mcp
86- Added tab completion for shell commands in bash mode
87- PreToolUse hooks can now modify tool inputs
88- Press Ctrl-G to edit your prompt in your system's configured text editor
89- Fixes for bash permission checks with environment variables in the command
90
91## 2.0.9
92
93- Fix regression where bash backgrounding stopped working
94
95## 2.0.8
96
97- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
98- IDE: Add drag-and-drop support for files and folders in chat
99- /context: Fix counting for thinking blocks
100- Improve message rendering for users with light themes on dark terminals
101- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
102
103## 2.0.5
104
105- IDE: Fix IME unintended message submission with Enter and Tab
106- IDE: Add "Open in Terminal" link in login screen
107- Fix unhandled OAuth expiration 401 API errors
108- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
109
110## 2.0.1
111
112- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
113- Various bug fixes and presentation improvements
114
115## 2.0.0
116
117- New native VS Code extension
118- Fresh coat of paint throughout the whole app
119- /rewind a conversation to undo code changes
120- /usage command to see plan limits
121- Tab to toggle thinking (sticky across sessions)
122- Ctrl-R to search history
123- Unshipped claude config command
124- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
125- SDK: The Claude Code SDK is now the Claude Agent SDK
126- Add subagents dynamically with `--agents` flag
127
128## 1.0.126
129
130- Enable /context command for Bedrock and Vertex
131- Add mTLS support for HTTP-based OpenTelemetry exporters
132
133## 1.0.124
134
135- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
136- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
137- No longer inform Claude of the list of allowed tools when permission is denied
138- Fixed security vulnerability in Bash tool permission checks
139- Improved VSCode extension performance for large files
140
141## 1.0.123
142
143- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
144- Fixed thinking mode triggering on negation phrases like "don't think"
145- Fixed rendering performance degradation during token streaming
146- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
147- Enhanced BashTool environment snapshot logging
148- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
149- Migrated --debug logging to a file, to enable easy tailing & filtering
150
151## 1.0.120
152
153- Fix input lag during typing, especially noticeable with large prompts
154- Improved VSCode extension command registry and sessions dialog user experience
155- Enhanced sessions dialog responsiveness and visual feedback
156- Fixed IDE compatibility issue by removing worktree support check
157- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
158
159## 1.0.119
160
161- Fix Windows issue where process visually freezes on entering interactive mode
162- Support dynamic headers for MCP servers via headersHelper configuration
163- Fix thinking mode not working in headless sessions
164- Fix slash commands now properly update allowed tools instead of replacing them
165
166## 1.0.117
167
168- Add Ctrl-R history search to recall previous commands like bash/zsh
169- Fix input lag while typing, especially on Windows
170- Add sed command to auto-allowed commands in acceptEdits mode
171- Fix Windows PATH comparison to be case-insensitive for drive letters
172- Add permissions management hint to /add-dir output
173
174## 1.0.115
175
176- Improve thinking mode display with enhanced visual effects
177- Type /t to temporarily disable thinking mode in your prompt
178- Improve path validation for glob and grep tools
179- Show condensed output for post-tool hooks to reduce visual clutter
180- Fix visual feedback when loading state completes
181- Improve UI consistency for permission request dialogs
182
183## 1.0.113
184
185- Deprecated piped input in interactive mode
186- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
187
188## 1.0.112
189
190- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
191- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
192- Hooks: Added systemMessage support for SessionEnd hooks
193- Added `spinnerTipsEnabled` setting to disable spinner tips
194- IDE: Various improvements and bug fixes
195
196## 1.0.111
197
198- /model now validates provided model names
199- Fixed Bash tool crashes caused by malformed shell syntax parsing
200
201## 1.0.110
202
203- /terminal-setup command now supports WezTerm
204- MCP: OAuth tokens now proactively refresh before expiration
205- Fixed reliability issues with background Bash processes
206
207## 1.0.109
208
209- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
210
211## 1.0.106
212
213- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
214
215## 1.0.97
216
217- Settings: /doctor now validates permission rule syntax and suggests corrections
218
219## 1.0.94
220
221- Vertex: add support for global endpoints for supported models
222- /memory command now allows direct editing of all imported memory files
223- SDK: Add custom tools as callbacks
224- Added /todos command to list current todo items
225
226## 1.0.93
227
228- Windows: Add alt + v shortcut for pasting images from clipboard
229- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
230
231## 1.0.90
232
233- Settings file changes take effect immediately - no restart required
234
235## 1.0.88
236
237- Fixed issue causing "OAuth authentication is currently not supported"
238- Status line input now includes `exceeds_200k_tokens`
239- Fixed incorrect usage tracking in /cost.
240- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
241- Bedrock: Updated default Sonnet model to Sonnet 4
242
243## 1.0.86
244
245- Added /context to help users self-serve debug context issues
246- SDK: Added UUID support for all SDK messages
247- SDK: Added `--replay-user-messages` to replay user messages back to stdout
248
249## 1.0.85
250
251- Status line input now includes session cost info
252- Hooks: Introduced SessionEnd hook
253
254## 1.0.84
255
256- Fix tool_use/tool_result id mismatch error when network is unstable
257- Fix Claude sometimes ignoring real-time steering when wrapping up a task
258- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
259- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
260
261## 1.0.83
262
263- @-mention: Support files with spaces in path
264- New shimmering spinner
265
266## 1.0.82
267
268- SDK: Add request cancellation support
269- SDK: New additionalDirectories option to search custom paths, improved slash command processing
270- Settings: Validation prevents invalid fields in .claude/settings.json files
271- MCP: Improve tool name consistency
272- Bash: Fix crash when Claude tries to automatically read large files
273
274## 1.0.81
275
276- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
277- Agents: Fix custom agent loading when agent files are unparsable
278
279## 1.0.80
280
281- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
282
283## 1.0.77
284
285- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
286- SDK: Add session support and permission denial tracking
287- Fix token limit errors in conversation summarization
288- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
289
290## 1.0.73
291
292- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
293- MCP: Press Esc to cancel OAuth authentication flows
294- Bash: Improved command validation and reduced false security warnings
295- UI: Enhanced spinner animations and status line visual hierarchy
296- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
297
298## 1.0.72
299
300- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
301
302## 1.0.71
303
304- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
305- Customizable status line: add your terminal prompt to Claude Code with /statusline
306
307## 1.0.70
308
309- Performance: Optimized message rendering for better performance with large contexts
310- Windows: Fixed native file search, ripgrep, and subagent functionality
311- Added support for @-mentions in slash command arguments
312
313## 1.0.69
314
315- Upgraded Opus to version 4.1
316
317## 1.0.68
318
319- Fix incorrect model names being used for certain commands like `/pr-comments`
320- 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.
321- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
322- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
323- SDK: Added canUseTool callback support for tool confirmation
324- Added `disableAllHooks` setting
325- Improved file suggestions performance in large repos
326
327## 1.0.65
328
329- IDE: Fixed connection stability issues and error handling for diagnostics
330- Windows: Fixed shell environment setup for users without .bashrc files
331
332## 1.0.64
333
334- Agents: Added model customization support - you can now specify which model an agent should use
335- Agents: Fixed unintended access to the recursive agent tool
336- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
337- SDK: Fixed user input tracking across multi-turn conversations
338- Added hidden files to file search and @-mention suggestions
339
340## 1.0.63
341
342- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
343
344## 1.0.62
345
346- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
347- Hooks: Added SessionStart hook for new session initialization
348- /add-dir command now supports typeahead for directory paths
349- Improved network connectivity check reliability
350
351## 1.0.61
352
353- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
354- Settings: Added `--settings` flag to load settings from a JSON file
355- Settings: Fixed resolution of settings files paths that are symlinks
356- OTEL: Fixed reporting of wrong organization after authentication changes
357- Slash commands: Fixed permissions checking for allowed-tools with Bash
358- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
359- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
360- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
361
362## 1.0.60
363
364- You can now create custom subagents for specialized tasks! Run /agents to get started
365
366## 1.0.59
367
368- SDK: Added tool confirmation support with canUseTool callback
369- SDK: Allow specifying env for spawned process
370- Hooks: Exposed PermissionDecision to hooks (including "ask")
371- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
372- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
373
374## 1.0.58
375
376- Added support for reading PDFs
377- MCP: Improved server health status display in 'claude mcp list'
378- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
379
380## 1.0.57
381
382- Added support for specifying a model in slash commands
383- Improved permission messages to help Claude understand allowed tools
384- Fix: Remove trailing newlines from bash output in terminal wrapping
385
386## 1.0.56
387
388- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
389- Fixes for WSL IDE detection
390- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
391
392## 1.0.55
393
394- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
395- Windows: fixed Ctrl+Z crash
396- SDK: Added ability to capture error logging
397- Add --system-prompt-file option to override system prompt in print mode
398
399## 1.0.54
400
401- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
402- Custom slash commands: Added argument-hint to frontmatter
403- Windows: OAuth uses port 45454 and properly constructs browser URL
404- Windows: mode switching now uses alt + m, and plan mode renders properly
405- Shell: Switch to in-memory shell snapshot to fix file-related errors
406
407## 1.0.53
408
409- Updated @-mention file truncation from 100 lines to 2000 lines
410- 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).
411
412## 1.0.52
413
414- Added support for MCP server instructions
415
416## 1.0.51
417
418- Added support for native Windows (requires Git for Windows)
419- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
420- Settings: /doctor can now help you identify and fix invalid setting files
421- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
422- Increased auto-compact warning threshold from 60% to 80%
423- Fixed an issue with handling user directories with spaces for shell snapshots
424- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
425- Custom slash commands: Fixed user-level commands in subdirectories
426- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
427
428## 1.0.48
429
430- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
431- Added progress messages to Bash tool based on the last 5 lines of command output
432- Added expanding variables support for MCP server configuration
433- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
434- Improved IDE extension path handling when Claude Code runs in WSL
435- Hooks: Added a PreCompact hook
436- Vim mode: Added c, f/F, t/T
437
438## 1.0.45
439
440- Redesigned Search (Grep) tool with new tool input parameters and features
441- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
442- Fixed config file corruption issue by enforcing atomic writes
443- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
444- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
445- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
446
447## 1.0.44
448
449- New /export command lets you quickly export a conversation for sharing
450- MCP: resource_link tool results are now supported
451- MCP: tool annotations and tool titles now display in /mcp view
452- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
453
454## 1.0.43
455
456- Fixed a bug where the theme selector was saving excessively
457- Hooks: Added EPIPE system error handling
458
459## 1.0.42
460
461- Added tilde (`~`) expansion support to `/add-dir` command
462
463## 1.0.41
464
465- Hooks: Split Stop hook triggering into Stop and SubagentStop
466- Hooks: Enabled optional timeout configuration for each command
467- Hooks: Added "hook_event_name" to hook input
468- Fixed a bug where MCP tools would display twice in tool list
469- New tool parameters JSON for Bash tool in `tool_decision` event
470
471## 1.0.40
472
473- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
474
475## 1.0.39
476
477- New Active Time metric in OpenTelemetry logging
478
479## 1.0.38
480
481- 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
482
483## 1.0.37
484
485- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
486
487## 1.0.36
488
489- Web search now takes today's date into context
490- Fixed a bug where stdio MCP servers were not terminating properly on exit
491
492## 1.0.35
493
494- Added support for MCP OAuth Authorization Server discovery
495
496## 1.0.34
497
498- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
499
500## 1.0.33
501
502- Improved logging functionality with session ID support
503- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
504- Improvements to plan mode
505
506## 1.0.32
507
508- Updated loopback config for litellm
509- Added forceLoginMethod setting to bypass login selection screen
510
511## 1.0.31
512
513- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
514
515## 1.0.30
516
517- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
518- Improved file path autocomplete with filename matching
519- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
520- Enhanced jq regex support for complex filters with pipes and select
521
522## 1.0.29
523
524- Improved CJK character support in cursor navigation and rendering
525
526## 1.0.28
527
528- Slash commands: Fix selector display during history navigation
529- Resizes images before upload to prevent API size limit errors
530- Added XDG_CONFIG_HOME support to configuration directory
531- Performance optimizations for memory usage
532- New attributes (terminal.type, language) in OpenTelemetry logging
533
534## 1.0.27
535
536- Streamable HTTP MCP servers are now supported
537- Remote MCP servers (SSE and HTTP) now support OAuth
538- MCP resources can now be @-mentioned
539- /resume slash command to switch conversations within Claude Code
540
541## 1.0.25
542
543- Slash commands: moved "project" and "user" prefixes to descriptions
544- Slash commands: improved reliability for command discovery
545- Improved support for Ghostty
546- Improved web search reliability
547
548## 1.0.24
549
550- Improved /mcp output
551- Fixed a bug where settings arrays got overwritten instead of merged
552
553## 1.0.23
554
555- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
556- Released Python SDK: pip install claude-code-sdk to get started
557
558## 1.0.22
559
560- SDK: Renamed `total_cost` to `total_cost_usd`
561
562## 1.0.21
563
564- Improved editing of files with tab-based indentation
565- Fix for tool_use without matching tool_result errors
566- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
567
568## 1.0.18
569
570- Added --add-dir CLI argument for specifying additional working directories
571- Added streaming input support without require -p flag
572- Improved startup performance and session storage performance
573- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
574- Added detailed MCP server tools display (/mcp)
575- MCP authentication and permission improvements
576- Added auto-reconnection for MCP SSE connections on disconnect
577- Fixed issue where pasted content was lost when dialogs appeared
578
579## 1.0.17
580
581- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
582- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
583- MCP server list UI improvements
584- Update Claude Code process title to display "claude" instead of "node"
585
586## 1.0.11
587
588- Claude Code can now also be used with a Claude Pro subscription
589- Added /upgrade for smoother switching to Claude Max plans
590- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
591- Improved shell configuration error handling
592- Improved todo list handling during compaction
593
594## 1.0.10
595
596- Added markdown table support
597- Improved streaming performance
598
599## 1.0.8
600
601- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
602- Increased default otel interval from 1s -> 5s
603- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
604- Fixed a regression where search tools unnecessarily asked for permissions
605- Added support for triggering thinking non-English languages
606- Improved compacting UI
607
608## 1.0.7
609
610- Renamed /allowed-tools -> /permissions
611- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
612- Deprecated claude config commands in favor of editing settings.json
613- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
614- Improved error handling for /install-github-app
615- Bugfixes, UI polish, and tool reliability improvements
616
617## 1.0.6
618
619- Improved edit reliability for tab-indented files
620- Respect CLAUDE_CONFIG_DIR everywhere
621- Reduced unnecessary tool permission prompts
622- Added support for symlinks in @file typeahead
623- Bugfixes, UI polish, and tool reliability improvements
624
625## 1.0.4
626
627- Fixed a bug where MCP tool errors weren't being parsed correctly
628
629## 1.0.1
630
631- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
632- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
633- Updated documentation links and OAuth process descriptions
634
635## 1.0.0
636
637- Claude Code is now generally available
638- Introducing Sonnet 4 and Opus 4 models
639
640## 0.2.125
641
642- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
643- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
644
645## 0.2.117
646
647- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
648- Introduced settings.cleanupPeriodDays
649- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
650- Introduced --debug mode
651
652## 0.2.108
653
654- You can now send messages to Claude while it works to steer Claude in real-time
655- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
656- Fixed a bug where thinking was not working in -p mode
657- Fixed a regression in /cost reporting
658- Deprecated MCP wizard interface in favor of other MCP commands
659- Lots of other bugfixes and improvements
660
661## 0.2.107
662
663- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
664
665## 0.2.106
666
667- MCP SSE server configs can now specify custom headers
668- Fixed a bug where MCP permission prompt didn't always show correctly
669
670## 0.2.105
671
672- Claude can now search the web
673- Moved system & account status to /status
674- Added word movement keybindings for Vim
675- Improved latency for startup, todo tool, and file edits
676
677## 0.2.102
678
679- Improved thinking triggering reliability
680- Improved @mention reliability for images and folders
681- You can now paste multiple large chunks into one prompt
682
683## 0.2.100
684
685- Fixed a crash caused by a stack overflow error
686- Made db storage optional; missing db support disables --continue and --resume
687
688## 0.2.98
689
690- Fixed an issue where auto-compact was running twice
691
692## 0.2.96
693
694- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
695
696## 0.2.93
697
698- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
699- Claude now has access to a Todo list that helps it stay on track and be more organized
700
701## 0.2.82
702
703- Added support for --disallowedTools
704- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
705
706## 0.2.75
707
708- Hit Enter to queue up additional messages while Claude is working
709- Drag in or copy/paste image files directly into the prompt
710- @-mention files to directly add them to context
711- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
712- Improved performance for filename auto-complete
713
714## 0.2.74
715
716- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
717- Task tool can now perform writes and run bash commands
718
719## 0.2.72
720
721- Updated spinner to indicate tokens loaded and tool usage
722
723## 0.2.70
724
725- Network commands like curl are now available for Claude to use
726- Claude can now run multiple web queries in parallel
727- Pressing ESC once immediately interrupts Claude in Auto-accept mode
728
729## 0.2.69
730
731- Fixed UI glitches with improved Select component behavior
732- Enhanced terminal output display with better text truncation logic
733
734## 0.2.67
735
736- Shared project permission rules can be saved in .claude/settings.json
737
738## 0.2.66
739
740- Print mode (-p) now supports streaming output via --output-format=stream-json
741- Fixed issue where pasting could trigger memory or bash mode unexpectedly
742
743## 0.2.63
744
745- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
746
747## 0.2.61
748
749- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
750- Enhanced image detection for more reliable clipboard paste functionality
751- Fixed an issue where ESC key could crash the conversation history selector
752
753## 0.2.59
754
755- Copy+paste images directly into your prompt
756- Improved progress indicators for bash and fetch tools
757- Bugfixes for non-interactive mode (-p)
758
759## 0.2.54
760
761- Quickly add to Memory by starting your message with '#'
762- Press ctrl+r to see full output for long tool results
763- Added support for MCP SSE transport
764
765## 0.2.53
766
767- New web fetch tool lets Claude view URLs that you paste in
768- Fixed a bug with JPEG detection
769
770## 0.2.50
771
772- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
773
774## 0.2.49
775
776- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
777
778## 0.2.47
779
780- Press Tab to auto-complete file and folder names
781- Press Shift + Tab to toggle auto-accept for file edits
782- Automatic conversation compaction for infinite conversation length (toggle with /config)
783
784## 0.2.44
785
786- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
787
788## 0.2.41
789
790- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
791- MCP server startup no longer blocks the app from starting up
792
793## 0.2.37
794
795- New /release-notes command lets you view release notes at any time
796- `claude config add/remove` commands now accept multiple values separated by commas or spaces
797
798## 0.2.36
799
800- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
801- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
802
803## 0.2.34
804
805- Vim bindings for text input - enable with /vim or /config
806
807## 0.2.32
808
809- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
810- Fix for some PersistentShell issues
811
812## 0.2.31
813
814- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
815- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
816
817## 0.2.30
818
819- Added ANSI color theme for better terminal compatibility
820- Fixed issue where slash command arguments weren't being sent properly
821- (Mac-only) API keys are now stored in macOS Keychain
822
823## 0.2.26
824
825- New /approved-tools command for managing tool permissions
826- Word-level diff display for improved code readability
827- Fuzzy matching for slash commands
828
829## 0.2.21
830
831- Fuzzy matching for /commands