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 b935da7 · on Oct 30, 2025 History
875 lines · 32.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.30
4
5- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
6- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
7- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
8- Added prompt-based stop hooks
9- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
10- Enabled SSE MCP servers on native build
11- Deprecated output styles. Review options in `/output-style` and use CLAUDE.md or plugins instead
12- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
13- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
14- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
15- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
16- Fixed bug that saved session permissions to local settings
17- Fixed MCP tools not being available to sub-agents
18- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
19- Fixed delay when navigating through typeahead suggestions with arrow keys
20- VSCode: Restored selection indicator in input footer showing current file or code selection status
21
22## 2.0.28
23
24- Plan mode: introduced new Plan subagent
25- Subagents: claude can now choose to resume subagents
26- Subagents: claude can dynamically choose the model used by its subagents
27- SDK: added --max-budget-usd flag
28- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
29- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
30- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
31- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
32- Various other bug fixes
33
34## 2.0.27
35
36- New UI for permission prompts
37- Added current branch filtering and search to session resume screen for easier navigation
38- Fixed directory @-mention causing "No assistant message found" error
39- VSCode Extension: Add config setting to include .gitignored files in file searches
40- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
41
42## 2.0.25
43
44- 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
45
46## 2.0.24
47
48- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
49- Claude Code Web: Support for Web -> CLI teleport
50- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
51- Bedrock: Display awsAuthRefresh output when auth is required
52
53## 2.0.22
54
55- Fixed content layout shift when scrolling through slash commands
56- IDE: Add toggle to enable/disable thinking.
57- Fix bug causing duplicate permission prompts with parallel tool calls
58- Add support for enterprise managed MCP allowlist and denylist
59
60## 2.0.21
61
62- Support MCP `structuredContent` field in tool responses
63- Added an interactive question tool
64- Claude will now ask you questions more often in plan mode
65- Added Haiku 4.5 as a model option for Pro users
66- Fixed an issue where queued commands don't have access to previous messages' output
67
68## 2.0.20
69
70- Added support for Claude Skills
71
72## 2.0.19
73
74- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
75- Fixed a bug where Haiku was unnecessarily called in print mode
76
77## 2.0.17
78
79- Added Haiku 4.5 to model selector!
80- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
81- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
82- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
83- OTEL: support HTTP_PROXY and HTTPS_PROXY
84- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
85
86## 2.0.15
87
88- Fixed bug with resuming where previously created files needed to be read again before writing
89- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
90
91## 2.0.14
92
93- Fix @-mentioning MCP servers to toggle them on/off
94- Improve permission checks for bash with inline env vars
95- Fix ultrathink + thinking toggle
96- Reduce unnecessary logins
97- Document --system-prompt
98- Several improvements to rendering
99- Plugins UI polish
100
101## 2.0.13
102
103- Fixed `/plugin` not working on native build
104
105## 2.0.12
106
107- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
108- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
109- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
110- `/plugin validate` command for validating plugin structure and configuration
111- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
112- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
113- Comprehensive error messages and diagnostics via `/doctor` command
114- Avoid flickering in `/model` selector
115- Improvements to `/help`
116- Avoid mentioning hooks in `/resume` summaries
117- Changes to the "verbose" setting in `/config` now persist across sessions
118
119## 2.0.11
120
121- Reduced system prompt size by 1.4k tokens
122- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
123- Fixed Opus fallback rate limit errors appearing incorrectly
124- Fixed /add-dir command selecting wrong default tab
125
126## 2.0.10
127
128- Rewrote terminal renderer for buttery smooth UI
129- Enable/disable MCP servers by @mentioning, or in /mcp
130- Added tab completion for shell commands in bash mode
131- PreToolUse hooks can now modify tool inputs
132- Press Ctrl-G to edit your prompt in your system's configured text editor
133- Fixes for bash permission checks with environment variables in the command
134
135## 2.0.9
136
137- Fix regression where bash backgrounding stopped working
138
139## 2.0.8
140
141- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
142- IDE: Add drag-and-drop support for files and folders in chat
143- /context: Fix counting for thinking blocks
144- Improve message rendering for users with light themes on dark terminals
145- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
146
147## 2.0.5
148
149- IDE: Fix IME unintended message submission with Enter and Tab
150- IDE: Add "Open in Terminal" link in login screen
151- Fix unhandled OAuth expiration 401 API errors
152- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
153
154## 2.0.1
155
156- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
157- Various bug fixes and presentation improvements
158
159## 2.0.0
160
161- New native VS Code extension
162- Fresh coat of paint throughout the whole app
163- /rewind a conversation to undo code changes
164- /usage command to see plan limits
165- Tab to toggle thinking (sticky across sessions)
166- Ctrl-R to search history
167- Unshipped claude config command
168- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
169- SDK: The Claude Code SDK is now the Claude Agent SDK
170- Add subagents dynamically with `--agents` flag
171
172## 1.0.126
173
174- Enable /context command for Bedrock and Vertex
175- Add mTLS support for HTTP-based OpenTelemetry exporters
176
177## 1.0.124
178
179- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
180- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
181- No longer inform Claude of the list of allowed tools when permission is denied
182- Fixed security vulnerability in Bash tool permission checks
183- Improved VSCode extension performance for large files
184
185## 1.0.123
186
187- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
188- Fixed thinking mode triggering on negation phrases like "don't think"
189- Fixed rendering performance degradation during token streaming
190- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
191- Enhanced BashTool environment snapshot logging
192- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
193- Migrated --debug logging to a file, to enable easy tailing & filtering
194
195## 1.0.120
196
197- Fix input lag during typing, especially noticeable with large prompts
198- Improved VSCode extension command registry and sessions dialog user experience
199- Enhanced sessions dialog responsiveness and visual feedback
200- Fixed IDE compatibility issue by removing worktree support check
201- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
202
203## 1.0.119
204
205- Fix Windows issue where process visually freezes on entering interactive mode
206- Support dynamic headers for MCP servers via headersHelper configuration
207- Fix thinking mode not working in headless sessions
208- Fix slash commands now properly update allowed tools instead of replacing them
209
210## 1.0.117
211
212- Add Ctrl-R history search to recall previous commands like bash/zsh
213- Fix input lag while typing, especially on Windows
214- Add sed command to auto-allowed commands in acceptEdits mode
215- Fix Windows PATH comparison to be case-insensitive for drive letters
216- Add permissions management hint to /add-dir output
217
218## 1.0.115
219
220- Improve thinking mode display with enhanced visual effects
221- Type /t to temporarily disable thinking mode in your prompt
222- Improve path validation for glob and grep tools
223- Show condensed output for post-tool hooks to reduce visual clutter
224- Fix visual feedback when loading state completes
225- Improve UI consistency for permission request dialogs
226
227## 1.0.113
228
229- Deprecated piped input in interactive mode
230- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
231
232## 1.0.112
233
234- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
235- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
236- Hooks: Added systemMessage support for SessionEnd hooks
237- Added `spinnerTipsEnabled` setting to disable spinner tips
238- IDE: Various improvements and bug fixes
239
240## 1.0.111
241
242- /model now validates provided model names
243- Fixed Bash tool crashes caused by malformed shell syntax parsing
244
245## 1.0.110
246
247- /terminal-setup command now supports WezTerm
248- MCP: OAuth tokens now proactively refresh before expiration
249- Fixed reliability issues with background Bash processes
250
251## 1.0.109
252
253- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
254
255## 1.0.106
256
257- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
258
259## 1.0.97
260
261- Settings: /doctor now validates permission rule syntax and suggests corrections
262
263## 1.0.94
264
265- Vertex: add support for global endpoints for supported models
266- /memory command now allows direct editing of all imported memory files
267- SDK: Add custom tools as callbacks
268- Added /todos command to list current todo items
269
270## 1.0.93
271
272- Windows: Add alt + v shortcut for pasting images from clipboard
273- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
274
275## 1.0.90
276
277- Settings file changes take effect immediately - no restart required
278
279## 1.0.88
280
281- Fixed issue causing "OAuth authentication is currently not supported"
282- Status line input now includes `exceeds_200k_tokens`
283- Fixed incorrect usage tracking in /cost.
284- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
285- Bedrock: Updated default Sonnet model to Sonnet 4
286
287## 1.0.86
288
289- Added /context to help users self-serve debug context issues
290- SDK: Added UUID support for all SDK messages
291- SDK: Added `--replay-user-messages` to replay user messages back to stdout
292
293## 1.0.85
294
295- Status line input now includes session cost info
296- Hooks: Introduced SessionEnd hook
297
298## 1.0.84
299
300- Fix tool_use/tool_result id mismatch error when network is unstable
301- Fix Claude sometimes ignoring real-time steering when wrapping up a task
302- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
303- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
304
305## 1.0.83
306
307- @-mention: Support files with spaces in path
308- New shimmering spinner
309
310## 1.0.82
311
312- SDK: Add request cancellation support
313- SDK: New additionalDirectories option to search custom paths, improved slash command processing
314- Settings: Validation prevents invalid fields in .claude/settings.json files
315- MCP: Improve tool name consistency
316- Bash: Fix crash when Claude tries to automatically read large files
317
318## 1.0.81
319
320- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
321- Agents: Fix custom agent loading when agent files are unparsable
322
323## 1.0.80
324
325- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
326
327## 1.0.77
328
329- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
330- SDK: Add session support and permission denial tracking
331- Fix token limit errors in conversation summarization
332- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
333
334## 1.0.73
335
336- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
337- MCP: Press Esc to cancel OAuth authentication flows
338- Bash: Improved command validation and reduced false security warnings
339- UI: Enhanced spinner animations and status line visual hierarchy
340- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
341
342## 1.0.72
343
344- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
345
346## 1.0.71
347
348- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
349- Customizable status line: add your terminal prompt to Claude Code with /statusline
350
351## 1.0.70
352
353- Performance: Optimized message rendering for better performance with large contexts
354- Windows: Fixed native file search, ripgrep, and subagent functionality
355- Added support for @-mentions in slash command arguments
356
357## 1.0.69
358
359- Upgraded Opus to version 4.1
360
361## 1.0.68
362
363- Fix incorrect model names being used for certain commands like `/pr-comments`
364- 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.
365- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
366- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
367- SDK: Added canUseTool callback support for tool confirmation
368- Added `disableAllHooks` setting
369- Improved file suggestions performance in large repos
370
371## 1.0.65
372
373- IDE: Fixed connection stability issues and error handling for diagnostics
374- Windows: Fixed shell environment setup for users without .bashrc files
375
376## 1.0.64
377
378- Agents: Added model customization support - you can now specify which model an agent should use
379- Agents: Fixed unintended access to the recursive agent tool
380- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
381- SDK: Fixed user input tracking across multi-turn conversations
382- Added hidden files to file search and @-mention suggestions
383
384## 1.0.63
385
386- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
387
388## 1.0.62
389
390- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
391- Hooks: Added SessionStart hook for new session initialization
392- /add-dir command now supports typeahead for directory paths
393- Improved network connectivity check reliability
394
395## 1.0.61
396
397- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
398- Settings: Added `--settings` flag to load settings from a JSON file
399- Settings: Fixed resolution of settings files paths that are symlinks
400- OTEL: Fixed reporting of wrong organization after authentication changes
401- Slash commands: Fixed permissions checking for allowed-tools with Bash
402- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
403- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
404- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
405
406## 1.0.60
407
408- You can now create custom subagents for specialized tasks! Run /agents to get started
409
410## 1.0.59
411
412- SDK: Added tool confirmation support with canUseTool callback
413- SDK: Allow specifying env for spawned process
414- Hooks: Exposed PermissionDecision to hooks (including "ask")
415- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
416- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
417
418## 1.0.58
419
420- Added support for reading PDFs
421- MCP: Improved server health status display in 'claude mcp list'
422- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
423
424## 1.0.57
425
426- Added support for specifying a model in slash commands
427- Improved permission messages to help Claude understand allowed tools
428- Fix: Remove trailing newlines from bash output in terminal wrapping
429
430## 1.0.56
431
432- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
433- Fixes for WSL IDE detection
434- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
435
436## 1.0.55
437
438- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
439- Windows: fixed Ctrl+Z crash
440- SDK: Added ability to capture error logging
441- Add --system-prompt-file option to override system prompt in print mode
442
443## 1.0.54
444
445- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
446- Custom slash commands: Added argument-hint to frontmatter
447- Windows: OAuth uses port 45454 and properly constructs browser URL
448- Windows: mode switching now uses alt + m, and plan mode renders properly
449- Shell: Switch to in-memory shell snapshot to fix file-related errors
450
451## 1.0.53
452
453- Updated @-mention file truncation from 100 lines to 2000 lines
454- 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).
455
456## 1.0.52
457
458- Added support for MCP server instructions
459
460## 1.0.51
461
462- Added support for native Windows (requires Git for Windows)
463- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
464- Settings: /doctor can now help you identify and fix invalid setting files
465- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
466- Increased auto-compact warning threshold from 60% to 80%
467- Fixed an issue with handling user directories with spaces for shell snapshots
468- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
469- Custom slash commands: Fixed user-level commands in subdirectories
470- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
471
472## 1.0.48
473
474- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
475- Added progress messages to Bash tool based on the last 5 lines of command output
476- Added expanding variables support for MCP server configuration
477- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
478- Improved IDE extension path handling when Claude Code runs in WSL
479- Hooks: Added a PreCompact hook
480- Vim mode: Added c, f/F, t/T
481
482## 1.0.45
483
484- Redesigned Search (Grep) tool with new tool input parameters and features
485- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
486- Fixed config file corruption issue by enforcing atomic writes
487- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
488- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
489- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
490
491## 1.0.44
492
493- New /export command lets you quickly export a conversation for sharing
494- MCP: resource_link tool results are now supported
495- MCP: tool annotations and tool titles now display in /mcp view
496- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
497
498## 1.0.43
499
500- Fixed a bug where the theme selector was saving excessively
501- Hooks: Added EPIPE system error handling
502
503## 1.0.42
504
505- Added tilde (`~`) expansion support to `/add-dir` command
506
507## 1.0.41
508
509- Hooks: Split Stop hook triggering into Stop and SubagentStop
510- Hooks: Enabled optional timeout configuration for each command
511- Hooks: Added "hook_event_name" to hook input
512- Fixed a bug where MCP tools would display twice in tool list
513- New tool parameters JSON for Bash tool in `tool_decision` event
514
515## 1.0.40
516
517- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
518
519## 1.0.39
520
521- New Active Time metric in OpenTelemetry logging
522
523## 1.0.38
524
525- 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
526
527## 1.0.37
528
529- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
530
531## 1.0.36
532
533- Web search now takes today's date into context
534- Fixed a bug where stdio MCP servers were not terminating properly on exit
535
536## 1.0.35
537
538- Added support for MCP OAuth Authorization Server discovery
539
540## 1.0.34
541
542- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
543
544## 1.0.33
545
546- Improved logging functionality with session ID support
547- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
548- Improvements to plan mode
549
550## 1.0.32
551
552- Updated loopback config for litellm
553- Added forceLoginMethod setting to bypass login selection screen
554
555## 1.0.31
556
557- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
558
559## 1.0.30
560
561- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
562- Improved file path autocomplete with filename matching
563- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
564- Enhanced jq regex support for complex filters with pipes and select
565
566## 1.0.29
567
568- Improved CJK character support in cursor navigation and rendering
569
570## 1.0.28
571
572- Slash commands: Fix selector display during history navigation
573- Resizes images before upload to prevent API size limit errors
574- Added XDG_CONFIG_HOME support to configuration directory
575- Performance optimizations for memory usage
576- New attributes (terminal.type, language) in OpenTelemetry logging
577
578## 1.0.27
579
580- Streamable HTTP MCP servers are now supported
581- Remote MCP servers (SSE and HTTP) now support OAuth
582- MCP resources can now be @-mentioned
583- /resume slash command to switch conversations within Claude Code
584
585## 1.0.25
586
587- Slash commands: moved "project" and "user" prefixes to descriptions
588- Slash commands: improved reliability for command discovery
589- Improved support for Ghostty
590- Improved web search reliability
591
592## 1.0.24
593
594- Improved /mcp output
595- Fixed a bug where settings arrays got overwritten instead of merged
596
597## 1.0.23
598
599- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
600- Released Python SDK: pip install claude-code-sdk to get started
601
602## 1.0.22
603
604- SDK: Renamed `total_cost` to `total_cost_usd`
605
606## 1.0.21
607
608- Improved editing of files with tab-based indentation
609- Fix for tool_use without matching tool_result errors
610- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
611
612## 1.0.18
613
614- Added --add-dir CLI argument for specifying additional working directories
615- Added streaming input support without require -p flag
616- Improved startup performance and session storage performance
617- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
618- Added detailed MCP server tools display (/mcp)
619- MCP authentication and permission improvements
620- Added auto-reconnection for MCP SSE connections on disconnect
621- Fixed issue where pasted content was lost when dialogs appeared
622
623## 1.0.17
624
625- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
626- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
627- MCP server list UI improvements
628- Update Claude Code process title to display "claude" instead of "node"
629
630## 1.0.11
631
632- Claude Code can now also be used with a Claude Pro subscription
633- Added /upgrade for smoother switching to Claude Max plans
634- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
635- Improved shell configuration error handling
636- Improved todo list handling during compaction
637
638## 1.0.10
639
640- Added markdown table support
641- Improved streaming performance
642
643## 1.0.8
644
645- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
646- Increased default otel interval from 1s -> 5s
647- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
648- Fixed a regression where search tools unnecessarily asked for permissions
649- Added support for triggering thinking non-English languages
650- Improved compacting UI
651
652## 1.0.7
653
654- Renamed /allowed-tools -> /permissions
655- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
656- Deprecated claude config commands in favor of editing settings.json
657- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
658- Improved error handling for /install-github-app
659- Bugfixes, UI polish, and tool reliability improvements
660
661## 1.0.6
662
663- Improved edit reliability for tab-indented files
664- Respect CLAUDE_CONFIG_DIR everywhere
665- Reduced unnecessary tool permission prompts
666- Added support for symlinks in @file typeahead
667- Bugfixes, UI polish, and tool reliability improvements
668
669## 1.0.4
670
671- Fixed a bug where MCP tool errors weren't being parsed correctly
672
673## 1.0.1
674
675- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
676- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
677- Updated documentation links and OAuth process descriptions
678
679## 1.0.0
680
681- Claude Code is now generally available
682- Introducing Sonnet 4 and Opus 4 models
683
684## 0.2.125
685
686- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
687- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
688
689## 0.2.117
690
691- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
692- Introduced settings.cleanupPeriodDays
693- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
694- Introduced --debug mode
695
696## 0.2.108
697
698- You can now send messages to Claude while it works to steer Claude in real-time
699- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
700- Fixed a bug where thinking was not working in -p mode
701- Fixed a regression in /cost reporting
702- Deprecated MCP wizard interface in favor of other MCP commands
703- Lots of other bugfixes and improvements
704
705## 0.2.107
706
707- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
708
709## 0.2.106
710
711- MCP SSE server configs can now specify custom headers
712- Fixed a bug where MCP permission prompt didn't always show correctly
713
714## 0.2.105
715
716- Claude can now search the web
717- Moved system & account status to /status
718- Added word movement keybindings for Vim
719- Improved latency for startup, todo tool, and file edits
720
721## 0.2.102
722
723- Improved thinking triggering reliability
724- Improved @mention reliability for images and folders
725- You can now paste multiple large chunks into one prompt
726
727## 0.2.100
728
729- Fixed a crash caused by a stack overflow error
730- Made db storage optional; missing db support disables --continue and --resume
731
732## 0.2.98
733
734- Fixed an issue where auto-compact was running twice
735
736## 0.2.96
737
738- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
739
740## 0.2.93
741
742- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
743- Claude now has access to a Todo list that helps it stay on track and be more organized
744
745## 0.2.82
746
747- Added support for --disallowedTools
748- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
749
750## 0.2.75
751
752- Hit Enter to queue up additional messages while Claude is working
753- Drag in or copy/paste image files directly into the prompt
754- @-mention files to directly add them to context
755- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
756- Improved performance for filename auto-complete
757
758## 0.2.74
759
760- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
761- Task tool can now perform writes and run bash commands
762
763## 0.2.72
764
765- Updated spinner to indicate tokens loaded and tool usage
766
767## 0.2.70
768
769- Network commands like curl are now available for Claude to use
770- Claude can now run multiple web queries in parallel
771- Pressing ESC once immediately interrupts Claude in Auto-accept mode
772
773## 0.2.69
774
775- Fixed UI glitches with improved Select component behavior
776- Enhanced terminal output display with better text truncation logic
777
778## 0.2.67
779
780- Shared project permission rules can be saved in .claude/settings.json
781
782## 0.2.66
783
784- Print mode (-p) now supports streaming output via --output-format=stream-json
785- Fixed issue where pasting could trigger memory or bash mode unexpectedly
786
787## 0.2.63
788
789- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
790
791## 0.2.61
792
793- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
794- Enhanced image detection for more reliable clipboard paste functionality
795- Fixed an issue where ESC key could crash the conversation history selector
796
797## 0.2.59
798
799- Copy+paste images directly into your prompt
800- Improved progress indicators for bash and fetch tools
801- Bugfixes for non-interactive mode (-p)
802
803## 0.2.54
804
805- Quickly add to Memory by starting your message with '#'
806- Press ctrl+r to see full output for long tool results
807- Added support for MCP SSE transport
808
809## 0.2.53
810
811- New web fetch tool lets Claude view URLs that you paste in
812- Fixed a bug with JPEG detection
813
814## 0.2.50
815
816- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
817
818## 0.2.49
819
820- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
821
822## 0.2.47
823
824- Press Tab to auto-complete file and folder names
825- Press Shift + Tab to toggle auto-accept for file edits
826- Automatic conversation compaction for infinite conversation length (toggle with /config)
827
828## 0.2.44
829
830- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
831
832## 0.2.41
833
834- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
835- MCP server startup no longer blocks the app from starting up
836
837## 0.2.37
838
839- New /release-notes command lets you view release notes at any time
840- `claude config add/remove` commands now accept multiple values separated by commas or spaces
841
842## 0.2.36
843
844- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
845- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
846
847## 0.2.34
848
849- Vim bindings for text input - enable with /vim or /config
850
851## 0.2.32
852
853- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
854- Fix for some PersistentShell issues
855
856## 0.2.31
857
858- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
859- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
860
861## 0.2.30
862
863- Added ANSI color theme for better terminal compatibility
864- Fixed issue where slash command arguments weren't being sent properly
865- (Mac-only) API keys are now stored in macOS Keychain
866
867## 0.2.26
868
869- New /approved-tools command for managing tool permissions
870- Word-level diff display for improved code readability
871- Fuzzy matching for slash commands
872
873## 0.2.21
874
875- Fuzzy matching for /commands