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 fb9d916 · on Oct 18, 2025 History
825 lines · 29.6 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.22
4
5- Fixed content layout shift when scrolling through slash commands
6- IDE: Add toggle to enable/disable thinking.
7- Fix bug causing duplicate permission prompts with parallel tool calls
8- Add support for enterprise managed MCP allowlist and denylist
9
10## 2.0.21
11
12- Support MCP `structuredContent` field in tool responses
13- Added an interactive question tool
14- Claude will now ask you questions more often in plan mode
15- Added Haiku 4.5 as a model option for Pro users
16- Fixed an issue where queued commands don't have access to previous messages' output
17
18## 2.0.20
19
20- Added support for Claude Skills
21
22## 2.0.19
23
24- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
25- Fixed a bug where Haiku was unnecessarily called in print mode
26
27## 2.0.17
28
29- Added Haiku 4.5 to model selector!
30- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
31- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
32- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
33- OTEL: support HTTP_PROXY and HTTPS_PROXY
34- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
35
36## 2.0.15
37
38- Fixed bug with resuming where previously created files needed to be read again before writing
39- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
40
41## 2.0.14
42
43- Fix @-mentioning MCP servers to toggle them on/off
44- Improve permission checks for bash with inline env vars
45- Fix ultrathink + thinking toggle
46- Reduce unnecessary logins
47- Document --system-prompt
48- Several improvements to rendering
49- Plugins UI polish
50
51## 2.0.13
52
53- Fixed `/plugin` not working on native build
54
55## 2.0.12
56
57- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
58- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
59- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
60- `/plugin validate` command for validating plugin structure and configuration
61- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
62- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
63- Comprehensive error messages and diagnostics via `/doctor` command
64- Avoid flickering in `/model` selector
65- Improvements to `/help`
66- Avoid mentioning hooks in `/resume` summaries
67- Changes to the "verbose" setting in `/config` now persist across sessions
68
69## 2.0.11
70
71- Reduced system prompt size by 1.4k tokens
72- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
73- Fixed Opus fallback rate limit errors appearing incorrectly
74- Fixed /add-dir command selecting wrong default tab
75
76## 2.0.10
77
78- Rewrote terminal renderer for buttery smooth UI
79- Enable/disable MCP servers by @mentioning, or in /mcp
80- Added tab completion for shell commands in bash mode
81- PreToolUse hooks can now modify tool inputs
82- Press Ctrl-G to edit your prompt in your system's configured text editor
83- Fixes for bash permission checks with environment variables in the command
84
85## 2.0.9
86
87- Fix regression where bash backgrounding stopped working
88
89## 2.0.8
90
91- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
92- IDE: Add drag-and-drop support for files and folders in chat
93- /context: Fix counting for thinking blocks
94- Improve message rendering for users with light themes on dark terminals
95- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
96
97## 2.0.5
98
99- IDE: Fix IME unintended message submission with Enter and Tab
100- IDE: Add "Open in Terminal" link in login screen
101- Fix unhandled OAuth expiration 401 API errors
102- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
103
104## 2.0.1
105
106- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
107- Various bug fixes and presentation improvements
108
109## 2.0.0
110
111- New native VS Code extension
112- Fresh coat of paint throughout the whole app
113- /rewind a conversation to undo code changes
114- /usage command to see plan limits
115- Tab to toggle thinking (sticky across sessions)
116- Ctrl-R to search history
117- Unshipped claude config command
118- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
119- SDK: The Claude Code SDK is now the Claude Agent SDK
120- Add subagents dynamically with `--agents` flag
121
122## 1.0.126
123
124- Enable /context command for Bedrock and Vertex
125- Add mTLS support for HTTP-based OpenTelemetry exporters
126
127## 1.0.124
128
129- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
130- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
131- No longer inform Claude of the list of allowed tools when permission is denied
132- Fixed security vulnerability in Bash tool permission checks
133- Improved VSCode extension performance for large files
134
135## 1.0.123
136
137- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
138- Fixed thinking mode triggering on negation phrases like "don't think"
139- Fixed rendering performance degradation during token streaming
140- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
141- Enhanced BashTool environment snapshot logging
142- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
143- Migrated --debug logging to a file, to enable easy tailing & filtering
144
145## 1.0.120
146
147- Fix input lag during typing, especially noticeable with large prompts
148- Improved VSCode extension command registry and sessions dialog user experience
149- Enhanced sessions dialog responsiveness and visual feedback
150- Fixed IDE compatibility issue by removing worktree support check
151- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
152
153## 1.0.119
154
155- Fix Windows issue where process visually freezes on entering interactive mode
156- Support dynamic headers for MCP servers via headersHelper configuration
157- Fix thinking mode not working in headless sessions
158- Fix slash commands now properly update allowed tools instead of replacing them
159
160## 1.0.117
161
162- Add Ctrl-R history search to recall previous commands like bash/zsh
163- Fix input lag while typing, especially on Windows
164- Add sed command to auto-allowed commands in acceptEdits mode
165- Fix Windows PATH comparison to be case-insensitive for drive letters
166- Add permissions management hint to /add-dir output
167
168## 1.0.115
169
170- Improve thinking mode display with enhanced visual effects
171- Type /t to temporarily disable thinking mode in your prompt
172- Improve path validation for glob and grep tools
173- Show condensed output for post-tool hooks to reduce visual clutter
174- Fix visual feedback when loading state completes
175- Improve UI consistency for permission request dialogs
176
177## 1.0.113
178
179- Deprecated piped input in interactive mode
180- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
181
182## 1.0.112
183
184- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
185- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
186- Hooks: Added systemMessage support for SessionEnd hooks
187- Added `spinnerTipsEnabled` setting to disable spinner tips
188- IDE: Various improvements and bug fixes
189
190## 1.0.111
191
192- /model now validates provided model names
193- Fixed Bash tool crashes caused by malformed shell syntax parsing
194
195## 1.0.110
196
197- /terminal-setup command now supports WezTerm
198- MCP: OAuth tokens now proactively refresh before expiration
199- Fixed reliability issues with background Bash processes
200
201## 1.0.109
202
203- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
204
205## 1.0.106
206
207- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
208
209## 1.0.97
210
211- Settings: /doctor now validates permission rule syntax and suggests corrections
212
213## 1.0.94
214
215- Vertex: add support for global endpoints for supported models
216- /memory command now allows direct editing of all imported memory files
217- SDK: Add custom tools as callbacks
218- Added /todos command to list current todo items
219
220## 1.0.93
221
222- Windows: Add alt + v shortcut for pasting images from clipboard
223- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
224
225## 1.0.90
226
227- Settings file changes take effect immediately - no restart required
228
229## 1.0.88
230
231- Fixed issue causing "OAuth authentication is currently not supported"
232- Status line input now includes `exceeds_200k_tokens`
233- Fixed incorrect usage tracking in /cost.
234- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
235- Bedrock: Updated default Sonnet model to Sonnet 4
236
237## 1.0.86
238
239- Added /context to help users self-serve debug context issues
240- SDK: Added UUID support for all SDK messages
241- SDK: Added `--replay-user-messages` to replay user messages back to stdout
242
243## 1.0.85
244
245- Status line input now includes session cost info
246- Hooks: Introduced SessionEnd hook
247
248## 1.0.84
249
250- Fix tool_use/tool_result id mismatch error when network is unstable
251- Fix Claude sometimes ignoring real-time steering when wrapping up a task
252- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
253- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
254
255## 1.0.83
256
257- @-mention: Support files with spaces in path
258- New shimmering spinner
259
260## 1.0.82
261
262- SDK: Add request cancellation support
263- SDK: New additionalDirectories option to search custom paths, improved slash command processing
264- Settings: Validation prevents invalid fields in .claude/settings.json files
265- MCP: Improve tool name consistency
266- Bash: Fix crash when Claude tries to automatically read large files
267
268## 1.0.81
269
270- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
271- Agents: Fix custom agent loading when agent files are unparsable
272
273## 1.0.80
274
275- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
276
277## 1.0.77
278
279- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
280- SDK: Add session support and permission denial tracking
281- Fix token limit errors in conversation summarization
282- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
283
284## 1.0.73
285
286- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
287- MCP: Press Esc to cancel OAuth authentication flows
288- Bash: Improved command validation and reduced false security warnings
289- UI: Enhanced spinner animations and status line visual hierarchy
290- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
291
292## 1.0.72
293
294- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
295
296## 1.0.71
297
298- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
299- Customizable status line: add your terminal prompt to Claude Code with /statusline
300
301## 1.0.70
302
303- Performance: Optimized message rendering for better performance with large contexts
304- Windows: Fixed native file search, ripgrep, and subagent functionality
305- Added support for @-mentions in slash command arguments
306
307## 1.0.69
308
309- Upgraded Opus to version 4.1
310
311## 1.0.68
312
313- Fix incorrect model names being used for certain commands like `/pr-comments`
314- 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.
315- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
316- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
317- SDK: Added canUseTool callback support for tool confirmation
318- Added `disableAllHooks` setting
319- Improved file suggestions performance in large repos
320
321## 1.0.65
322
323- IDE: Fixed connection stability issues and error handling for diagnostics
324- Windows: Fixed shell environment setup for users without .bashrc files
325
326## 1.0.64
327
328- Agents: Added model customization support - you can now specify which model an agent should use
329- Agents: Fixed unintended access to the recursive agent tool
330- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
331- SDK: Fixed user input tracking across multi-turn conversations
332- Added hidden files to file search and @-mention suggestions
333
334## 1.0.63
335
336- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
337
338## 1.0.62
339
340- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
341- Hooks: Added SessionStart hook for new session initialization
342- /add-dir command now supports typeahead for directory paths
343- Improved network connectivity check reliability
344
345## 1.0.61
346
347- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
348- Settings: Added `--settings` flag to load settings from a JSON file
349- Settings: Fixed resolution of settings files paths that are symlinks
350- OTEL: Fixed reporting of wrong organization after authentication changes
351- Slash commands: Fixed permissions checking for allowed-tools with Bash
352- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
353- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
354- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
355
356## 1.0.60
357
358- You can now create custom subagents for specialized tasks! Run /agents to get started
359
360## 1.0.59
361
362- SDK: Added tool confirmation support with canUseTool callback
363- SDK: Allow specifying env for spawned process
364- Hooks: Exposed PermissionDecision to hooks (including "ask")
365- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
366- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
367
368## 1.0.58
369
370- Added support for reading PDFs
371- MCP: Improved server health status display in 'claude mcp list'
372- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
373
374## 1.0.57
375
376- Added support for specifying a model in slash commands
377- Improved permission messages to help Claude understand allowed tools
378- Fix: Remove trailing newlines from bash output in terminal wrapping
379
380## 1.0.56
381
382- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
383- Fixes for WSL IDE detection
384- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
385
386## 1.0.55
387
388- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
389- Windows: fixed Ctrl+Z crash
390- SDK: Added ability to capture error logging
391- Add --system-prompt-file option to override system prompt in print mode
392
393## 1.0.54
394
395- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
396- Custom slash commands: Added argument-hint to frontmatter
397- Windows: OAuth uses port 45454 and properly constructs browser URL
398- Windows: mode switching now uses alt + m, and plan mode renders properly
399- Shell: Switch to in-memory shell snapshot to fix file-related errors
400
401## 1.0.53
402
403- Updated @-mention file truncation from 100 lines to 2000 lines
404- 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).
405
406## 1.0.52
407
408- Added support for MCP server instructions
409
410## 1.0.51
411
412- Added support for native Windows (requires Git for Windows)
413- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
414- Settings: /doctor can now help you identify and fix invalid setting files
415- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
416- Increased auto-compact warning threshold from 60% to 80%
417- Fixed an issue with handling user directories with spaces for shell snapshots
418- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
419- Custom slash commands: Fixed user-level commands in subdirectories
420- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
421
422## 1.0.48
423
424- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
425- Added progress messages to Bash tool based on the last 5 lines of command output
426- Added expanding variables support for MCP server configuration
427- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
428- Improved IDE extension path handling when Claude Code runs in WSL
429- Hooks: Added a PreCompact hook
430- Vim mode: Added c, f/F, t/T
431
432## 1.0.45
433
434- Redesigned Search (Grep) tool with new tool input parameters and features
435- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
436- Fixed config file corruption issue by enforcing atomic writes
437- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
438- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
439- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
440
441## 1.0.44
442
443- New /export command lets you quickly export a conversation for sharing
444- MCP: resource_link tool results are now supported
445- MCP: tool annotations and tool titles now display in /mcp view
446- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
447
448## 1.0.43
449
450- Fixed a bug where the theme selector was saving excessively
451- Hooks: Added EPIPE system error handling
452
453## 1.0.42
454
455- Added tilde (`~`) expansion support to `/add-dir` command
456
457## 1.0.41
458
459- Hooks: Split Stop hook triggering into Stop and SubagentStop
460- Hooks: Enabled optional timeout configuration for each command
461- Hooks: Added "hook_event_name" to hook input
462- Fixed a bug where MCP tools would display twice in tool list
463- New tool parameters JSON for Bash tool in `tool_decision` event
464
465## 1.0.40
466
467- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
468
469## 1.0.39
470
471- New Active Time metric in OpenTelemetry logging
472
473## 1.0.38
474
475- 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
476
477## 1.0.37
478
479- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
480
481## 1.0.36
482
483- Web search now takes today's date into context
484- Fixed a bug where stdio MCP servers were not terminating properly on exit
485
486## 1.0.35
487
488- Added support for MCP OAuth Authorization Server discovery
489
490## 1.0.34
491
492- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
493
494## 1.0.33
495
496- Improved logging functionality with session ID support
497- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
498- Improvements to plan mode
499
500## 1.0.32
501
502- Updated loopback config for litellm
503- Added forceLoginMethod setting to bypass login selection screen
504
505## 1.0.31
506
507- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
508
509## 1.0.30
510
511- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
512- Improved file path autocomplete with filename matching
513- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
514- Enhanced jq regex support for complex filters with pipes and select
515
516## 1.0.29
517
518- Improved CJK character support in cursor navigation and rendering
519
520## 1.0.28
521
522- Slash commands: Fix selector display during history navigation
523- Resizes images before upload to prevent API size limit errors
524- Added XDG_CONFIG_HOME support to configuration directory
525- Performance optimizations for memory usage
526- New attributes (terminal.type, language) in OpenTelemetry logging
527
528## 1.0.27
529
530- Streamable HTTP MCP servers are now supported
531- Remote MCP servers (SSE and HTTP) now support OAuth
532- MCP resources can now be @-mentioned
533- /resume slash command to switch conversations within Claude Code
534
535## 1.0.25
536
537- Slash commands: moved "project" and "user" prefixes to descriptions
538- Slash commands: improved reliability for command discovery
539- Improved support for Ghostty
540- Improved web search reliability
541
542## 1.0.24
543
544- Improved /mcp output
545- Fixed a bug where settings arrays got overwritten instead of merged
546
547## 1.0.23
548
549- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
550- Released Python SDK: pip install claude-code-sdk to get started
551
552## 1.0.22
553
554- SDK: Renamed `total_cost` to `total_cost_usd`
555
556## 1.0.21
557
558- Improved editing of files with tab-based indentation
559- Fix for tool_use without matching tool_result errors
560- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
561
562## 1.0.18
563
564- Added --add-dir CLI argument for specifying additional working directories
565- Added streaming input support without require -p flag
566- Improved startup performance and session storage performance
567- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
568- Added detailed MCP server tools display (/mcp)
569- MCP authentication and permission improvements
570- Added auto-reconnection for MCP SSE connections on disconnect
571- Fixed issue where pasted content was lost when dialogs appeared
572
573## 1.0.17
574
575- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
576- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
577- MCP server list UI improvements
578- Update Claude Code process title to display "claude" instead of "node"
579
580## 1.0.11
581
582- Claude Code can now also be used with a Claude Pro subscription
583- Added /upgrade for smoother switching to Claude Max plans
584- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
585- Improved shell configuration error handling
586- Improved todo list handling during compaction
587
588## 1.0.10
589
590- Added markdown table support
591- Improved streaming performance
592
593## 1.0.8
594
595- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
596- Increased default otel interval from 1s -> 5s
597- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
598- Fixed a regression where search tools unnecessarily asked for permissions
599- Added support for triggering thinking non-English languages
600- Improved compacting UI
601
602## 1.0.7
603
604- Renamed /allowed-tools -> /permissions
605- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
606- Deprecated claude config commands in favor of editing settings.json
607- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
608- Improved error handling for /install-github-app
609- Bugfixes, UI polish, and tool reliability improvements
610
611## 1.0.6
612
613- Improved edit reliability for tab-indented files
614- Respect CLAUDE_CONFIG_DIR everywhere
615- Reduced unnecessary tool permission prompts
616- Added support for symlinks in @file typeahead
617- Bugfixes, UI polish, and tool reliability improvements
618
619## 1.0.4
620
621- Fixed a bug where MCP tool errors weren't being parsed correctly
622
623## 1.0.1
624
625- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
626- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
627- Updated documentation links and OAuth process descriptions
628
629## 1.0.0
630
631- Claude Code is now generally available
632- Introducing Sonnet 4 and Opus 4 models
633
634## 0.2.125
635
636- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
637- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
638
639## 0.2.117
640
641- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
642- Introduced settings.cleanupPeriodDays
643- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
644- Introduced --debug mode
645
646## 0.2.108
647
648- You can now send messages to Claude while it works to steer Claude in real-time
649- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
650- Fixed a bug where thinking was not working in -p mode
651- Fixed a regression in /cost reporting
652- Deprecated MCP wizard interface in favor of other MCP commands
653- Lots of other bugfixes and improvements
654
655## 0.2.107
656
657- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
658
659## 0.2.106
660
661- MCP SSE server configs can now specify custom headers
662- Fixed a bug where MCP permission prompt didn't always show correctly
663
664## 0.2.105
665
666- Claude can now search the web
667- Moved system & account status to /status
668- Added word movement keybindings for Vim
669- Improved latency for startup, todo tool, and file edits
670
671## 0.2.102
672
673- Improved thinking triggering reliability
674- Improved @mention reliability for images and folders
675- You can now paste multiple large chunks into one prompt
676
677## 0.2.100
678
679- Fixed a crash caused by a stack overflow error
680- Made db storage optional; missing db support disables --continue and --resume
681
682## 0.2.98
683
684- Fixed an issue where auto-compact was running twice
685
686## 0.2.96
687
688- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
689
690## 0.2.93
691
692- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
693- Claude now has access to a Todo list that helps it stay on track and be more organized
694
695## 0.2.82
696
697- Added support for --disallowedTools
698- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
699
700## 0.2.75
701
702- Hit Enter to queue up additional messages while Claude is working
703- Drag in or copy/paste image files directly into the prompt
704- @-mention files to directly add them to context
705- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
706- Improved performance for filename auto-complete
707
708## 0.2.74
709
710- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
711- Task tool can now perform writes and run bash commands
712
713## 0.2.72
714
715- Updated spinner to indicate tokens loaded and tool usage
716
717## 0.2.70
718
719- Network commands like curl are now available for Claude to use
720- Claude can now run multiple web queries in parallel
721- Pressing ESC once immediately interrupts Claude in Auto-accept mode
722
723## 0.2.69
724
725- Fixed UI glitches with improved Select component behavior
726- Enhanced terminal output display with better text truncation logic
727
728## 0.2.67
729
730- Shared project permission rules can be saved in .claude/settings.json
731
732## 0.2.66
733
734- Print mode (-p) now supports streaming output via --output-format=stream-json
735- Fixed issue where pasting could trigger memory or bash mode unexpectedly
736
737## 0.2.63
738
739- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
740
741## 0.2.61
742
743- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
744- Enhanced image detection for more reliable clipboard paste functionality
745- Fixed an issue where ESC key could crash the conversation history selector
746
747## 0.2.59
748
749- Copy+paste images directly into your prompt
750- Improved progress indicators for bash and fetch tools
751- Bugfixes for non-interactive mode (-p)
752
753## 0.2.54
754
755- Quickly add to Memory by starting your message with '#'
756- Press ctrl+r to see full output for long tool results
757- Added support for MCP SSE transport
758
759## 0.2.53
760
761- New web fetch tool lets Claude view URLs that you paste in
762- Fixed a bug with JPEG detection
763
764## 0.2.50
765
766- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
767
768## 0.2.49
769
770- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
771
772## 0.2.47
773
774- Press Tab to auto-complete file and folder names
775- Press Shift + Tab to toggle auto-accept for file edits
776- Automatic conversation compaction for infinite conversation length (toggle with /config)
777
778## 0.2.44
779
780- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
781
782## 0.2.41
783
784- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
785- MCP server startup no longer blocks the app from starting up
786
787## 0.2.37
788
789- New /release-notes command lets you view release notes at any time
790- `claude config add/remove` commands now accept multiple values separated by commas or spaces
791
792## 0.2.36
793
794- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
795- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
796
797## 0.2.34
798
799- Vim bindings for text input - enable with /vim or /config
800
801## 0.2.32
802
803- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
804- Fix for some PersistentShell issues
805
806## 0.2.31
807
808- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
809- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
810
811## 0.2.30
812
813- Added ANSI color theme for better terminal compatibility
814- Fixed issue where slash command arguments weren't being sent properly
815- (Mac-only) API keys are now stored in macOS Keychain
816
817## 0.2.26
818
819- New /approved-tools command for managing tool permissions
820- Word-level diff display for improved code readability
821- Fuzzy matching for slash commands
822
823## 0.2.21
824
825- Fuzzy matching for /commands