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 ae411f8 · on Nov 1, 2025 History
884 lines · 33.1 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.31
4
5- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
6- Vertex: add support for Web Search on supported models
7- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
8- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
9- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
10- Fixed plugin uninstall not removing plugins
11
12## 2.0.30
13
14- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
15- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
16- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
17- Added prompt-based stop hooks
18- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
19- Enabled SSE MCP servers on native build
20- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
21- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
22- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
23- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
24- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
25- Fixed bug that saved session permissions to local settings
26- Fixed MCP tools not being available to sub-agents
27- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
28- Fixed delay when navigating through typeahead suggestions with arrow keys
29- VSCode: Restored selection indicator in input footer showing current file or code selection status
30
31## 2.0.28
32
33- Plan mode: introduced new Plan subagent
34- Subagents: claude can now choose to resume subagents
35- Subagents: claude can dynamically choose the model used by its subagents
36- SDK: added --max-budget-usd flag
37- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
38- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
39- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
40- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
41- Various other bug fixes
42
43## 2.0.27
44
45- New UI for permission prompts
46- Added current branch filtering and search to session resume screen for easier navigation
47- Fixed directory @-mention causing "No assistant message found" error
48- VSCode Extension: Add config setting to include .gitignored files in file searches
49- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
50
51## 2.0.25
52
53- 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
54
55## 2.0.24
56
57- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
58- Claude Code Web: Support for Web -> CLI teleport
59- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
60- Bedrock: Display awsAuthRefresh output when auth is required
61
62## 2.0.22
63
64- Fixed content layout shift when scrolling through slash commands
65- IDE: Add toggle to enable/disable thinking.
66- Fix bug causing duplicate permission prompts with parallel tool calls
67- Add support for enterprise managed MCP allowlist and denylist
68
69## 2.0.21
70
71- Support MCP `structuredContent` field in tool responses
72- Added an interactive question tool
73- Claude will now ask you questions more often in plan mode
74- Added Haiku 4.5 as a model option for Pro users
75- Fixed an issue where queued commands don't have access to previous messages' output
76
77## 2.0.20
78
79- Added support for Claude Skills
80
81## 2.0.19
82
83- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
84- Fixed a bug where Haiku was unnecessarily called in print mode
85
86## 2.0.17
87
88- Added Haiku 4.5 to model selector!
89- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
90- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
91- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
92- OTEL: support HTTP_PROXY and HTTPS_PROXY
93- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
94
95## 2.0.15
96
97- Fixed bug with resuming where previously created files needed to be read again before writing
98- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
99
100## 2.0.14
101
102- Fix @-mentioning MCP servers to toggle them on/off
103- Improve permission checks for bash with inline env vars
104- Fix ultrathink + thinking toggle
105- Reduce unnecessary logins
106- Document --system-prompt
107- Several improvements to rendering
108- Plugins UI polish
109
110## 2.0.13
111
112- Fixed `/plugin` not working on native build
113
114## 2.0.12
115
116- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
117- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
118- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
119- `/plugin validate` command for validating plugin structure and configuration
120- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
121- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
122- Comprehensive error messages and diagnostics via `/doctor` command
123- Avoid flickering in `/model` selector
124- Improvements to `/help`
125- Avoid mentioning hooks in `/resume` summaries
126- Changes to the "verbose" setting in `/config` now persist across sessions
127
128## 2.0.11
129
130- Reduced system prompt size by 1.4k tokens
131- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
132- Fixed Opus fallback rate limit errors appearing incorrectly
133- Fixed /add-dir command selecting wrong default tab
134
135## 2.0.10
136
137- Rewrote terminal renderer for buttery smooth UI
138- Enable/disable MCP servers by @mentioning, or in /mcp
139- Added tab completion for shell commands in bash mode
140- PreToolUse hooks can now modify tool inputs
141- Press Ctrl-G to edit your prompt in your system's configured text editor
142- Fixes for bash permission checks with environment variables in the command
143
144## 2.0.9
145
146- Fix regression where bash backgrounding stopped working
147
148## 2.0.8
149
150- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
151- IDE: Add drag-and-drop support for files and folders in chat
152- /context: Fix counting for thinking blocks
153- Improve message rendering for users with light themes on dark terminals
154- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
155
156## 2.0.5
157
158- IDE: Fix IME unintended message submission with Enter and Tab
159- IDE: Add "Open in Terminal" link in login screen
160- Fix unhandled OAuth expiration 401 API errors
161- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
162
163## 2.0.1
164
165- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
166- Various bug fixes and presentation improvements
167
168## 2.0.0
169
170- New native VS Code extension
171- Fresh coat of paint throughout the whole app
172- /rewind a conversation to undo code changes
173- /usage command to see plan limits
174- Tab to toggle thinking (sticky across sessions)
175- Ctrl-R to search history
176- Unshipped claude config command
177- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
178- SDK: The Claude Code SDK is now the Claude Agent SDK
179- Add subagents dynamically with `--agents` flag
180
181## 1.0.126
182
183- Enable /context command for Bedrock and Vertex
184- Add mTLS support for HTTP-based OpenTelemetry exporters
185
186## 1.0.124
187
188- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
189- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
190- No longer inform Claude of the list of allowed tools when permission is denied
191- Fixed security vulnerability in Bash tool permission checks
192- Improved VSCode extension performance for large files
193
194## 1.0.123
195
196- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
197- Fixed thinking mode triggering on negation phrases like "don't think"
198- Fixed rendering performance degradation during token streaming
199- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
200- Enhanced BashTool environment snapshot logging
201- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
202- Migrated --debug logging to a file, to enable easy tailing & filtering
203
204## 1.0.120
205
206- Fix input lag during typing, especially noticeable with large prompts
207- Improved VSCode extension command registry and sessions dialog user experience
208- Enhanced sessions dialog responsiveness and visual feedback
209- Fixed IDE compatibility issue by removing worktree support check
210- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
211
212## 1.0.119
213
214- Fix Windows issue where process visually freezes on entering interactive mode
215- Support dynamic headers for MCP servers via headersHelper configuration
216- Fix thinking mode not working in headless sessions
217- Fix slash commands now properly update allowed tools instead of replacing them
218
219## 1.0.117
220
221- Add Ctrl-R history search to recall previous commands like bash/zsh
222- Fix input lag while typing, especially on Windows
223- Add sed command to auto-allowed commands in acceptEdits mode
224- Fix Windows PATH comparison to be case-insensitive for drive letters
225- Add permissions management hint to /add-dir output
226
227## 1.0.115
228
229- Improve thinking mode display with enhanced visual effects
230- Type /t to temporarily disable thinking mode in your prompt
231- Improve path validation for glob and grep tools
232- Show condensed output for post-tool hooks to reduce visual clutter
233- Fix visual feedback when loading state completes
234- Improve UI consistency for permission request dialogs
235
236## 1.0.113
237
238- Deprecated piped input in interactive mode
239- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
240
241## 1.0.112
242
243- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
244- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
245- Hooks: Added systemMessage support for SessionEnd hooks
246- Added `spinnerTipsEnabled` setting to disable spinner tips
247- IDE: Various improvements and bug fixes
248
249## 1.0.111
250
251- /model now validates provided model names
252- Fixed Bash tool crashes caused by malformed shell syntax parsing
253
254## 1.0.110
255
256- /terminal-setup command now supports WezTerm
257- MCP: OAuth tokens now proactively refresh before expiration
258- Fixed reliability issues with background Bash processes
259
260## 1.0.109
261
262- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
263
264## 1.0.106
265
266- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
267
268## 1.0.97
269
270- Settings: /doctor now validates permission rule syntax and suggests corrections
271
272## 1.0.94
273
274- Vertex: add support for global endpoints for supported models
275- /memory command now allows direct editing of all imported memory files
276- SDK: Add custom tools as callbacks
277- Added /todos command to list current todo items
278
279## 1.0.93
280
281- Windows: Add alt + v shortcut for pasting images from clipboard
282- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
283
284## 1.0.90
285
286- Settings file changes take effect immediately - no restart required
287
288## 1.0.88
289
290- Fixed issue causing "OAuth authentication is currently not supported"
291- Status line input now includes `exceeds_200k_tokens`
292- Fixed incorrect usage tracking in /cost.
293- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
294- Bedrock: Updated default Sonnet model to Sonnet 4
295
296## 1.0.86
297
298- Added /context to help users self-serve debug context issues
299- SDK: Added UUID support for all SDK messages
300- SDK: Added `--replay-user-messages` to replay user messages back to stdout
301
302## 1.0.85
303
304- Status line input now includes session cost info
305- Hooks: Introduced SessionEnd hook
306
307## 1.0.84
308
309- Fix tool_use/tool_result id mismatch error when network is unstable
310- Fix Claude sometimes ignoring real-time steering when wrapping up a task
311- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
312- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
313
314## 1.0.83
315
316- @-mention: Support files with spaces in path
317- New shimmering spinner
318
319## 1.0.82
320
321- SDK: Add request cancellation support
322- SDK: New additionalDirectories option to search custom paths, improved slash command processing
323- Settings: Validation prevents invalid fields in .claude/settings.json files
324- MCP: Improve tool name consistency
325- Bash: Fix crash when Claude tries to automatically read large files
326
327## 1.0.81
328
329- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
330- Agents: Fix custom agent loading when agent files are unparsable
331
332## 1.0.80
333
334- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
335
336## 1.0.77
337
338- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
339- SDK: Add session support and permission denial tracking
340- Fix token limit errors in conversation summarization
341- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
342
343## 1.0.73
344
345- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
346- MCP: Press Esc to cancel OAuth authentication flows
347- Bash: Improved command validation and reduced false security warnings
348- UI: Enhanced spinner animations and status line visual hierarchy
349- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
350
351## 1.0.72
352
353- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
354
355## 1.0.71
356
357- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
358- Customizable status line: add your terminal prompt to Claude Code with /statusline
359
360## 1.0.70
361
362- Performance: Optimized message rendering for better performance with large contexts
363- Windows: Fixed native file search, ripgrep, and subagent functionality
364- Added support for @-mentions in slash command arguments
365
366## 1.0.69
367
368- Upgraded Opus to version 4.1
369
370## 1.0.68
371
372- Fix incorrect model names being used for certain commands like `/pr-comments`
373- 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.
374- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
375- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
376- SDK: Added canUseTool callback support for tool confirmation
377- Added `disableAllHooks` setting
378- Improved file suggestions performance in large repos
379
380## 1.0.65
381
382- IDE: Fixed connection stability issues and error handling for diagnostics
383- Windows: Fixed shell environment setup for users without .bashrc files
384
385## 1.0.64
386
387- Agents: Added model customization support - you can now specify which model an agent should use
388- Agents: Fixed unintended access to the recursive agent tool
389- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
390- SDK: Fixed user input tracking across multi-turn conversations
391- Added hidden files to file search and @-mention suggestions
392
393## 1.0.63
394
395- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
396
397## 1.0.62
398
399- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
400- Hooks: Added SessionStart hook for new session initialization
401- /add-dir command now supports typeahead for directory paths
402- Improved network connectivity check reliability
403
404## 1.0.61
405
406- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
407- Settings: Added `--settings` flag to load settings from a JSON file
408- Settings: Fixed resolution of settings files paths that are symlinks
409- OTEL: Fixed reporting of wrong organization after authentication changes
410- Slash commands: Fixed permissions checking for allowed-tools with Bash
411- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
412- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
413- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
414
415## 1.0.60
416
417- You can now create custom subagents for specialized tasks! Run /agents to get started
418
419## 1.0.59
420
421- SDK: Added tool confirmation support with canUseTool callback
422- SDK: Allow specifying env for spawned process
423- Hooks: Exposed PermissionDecision to hooks (including "ask")
424- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
425- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
426
427## 1.0.58
428
429- Added support for reading PDFs
430- MCP: Improved server health status display in 'claude mcp list'
431- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
432
433## 1.0.57
434
435- Added support for specifying a model in slash commands
436- Improved permission messages to help Claude understand allowed tools
437- Fix: Remove trailing newlines from bash output in terminal wrapping
438
439## 1.0.56
440
441- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
442- Fixes for WSL IDE detection
443- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
444
445## 1.0.55
446
447- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
448- Windows: fixed Ctrl+Z crash
449- SDK: Added ability to capture error logging
450- Add --system-prompt-file option to override system prompt in print mode
451
452## 1.0.54
453
454- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
455- Custom slash commands: Added argument-hint to frontmatter
456- Windows: OAuth uses port 45454 and properly constructs browser URL
457- Windows: mode switching now uses alt + m, and plan mode renders properly
458- Shell: Switch to in-memory shell snapshot to fix file-related errors
459
460## 1.0.53
461
462- Updated @-mention file truncation from 100 lines to 2000 lines
463- 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).
464
465## 1.0.52
466
467- Added support for MCP server instructions
468
469## 1.0.51
470
471- Added support for native Windows (requires Git for Windows)
472- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
473- Settings: /doctor can now help you identify and fix invalid setting files
474- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
475- Increased auto-compact warning threshold from 60% to 80%
476- Fixed an issue with handling user directories with spaces for shell snapshots
477- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
478- Custom slash commands: Fixed user-level commands in subdirectories
479- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
480
481## 1.0.48
482
483- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
484- Added progress messages to Bash tool based on the last 5 lines of command output
485- Added expanding variables support for MCP server configuration
486- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
487- Improved IDE extension path handling when Claude Code runs in WSL
488- Hooks: Added a PreCompact hook
489- Vim mode: Added c, f/F, t/T
490
491## 1.0.45
492
493- Redesigned Search (Grep) tool with new tool input parameters and features
494- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
495- Fixed config file corruption issue by enforcing atomic writes
496- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
497- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
498- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
499
500## 1.0.44
501
502- New /export command lets you quickly export a conversation for sharing
503- MCP: resource_link tool results are now supported
504- MCP: tool annotations and tool titles now display in /mcp view
505- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
506
507## 1.0.43
508
509- Fixed a bug where the theme selector was saving excessively
510- Hooks: Added EPIPE system error handling
511
512## 1.0.42
513
514- Added tilde (`~`) expansion support to `/add-dir` command
515
516## 1.0.41
517
518- Hooks: Split Stop hook triggering into Stop and SubagentStop
519- Hooks: Enabled optional timeout configuration for each command
520- Hooks: Added "hook_event_name" to hook input
521- Fixed a bug where MCP tools would display twice in tool list
522- New tool parameters JSON for Bash tool in `tool_decision` event
523
524## 1.0.40
525
526- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
527
528## 1.0.39
529
530- New Active Time metric in OpenTelemetry logging
531
532## 1.0.38
533
534- 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
535
536## 1.0.37
537
538- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
539
540## 1.0.36
541
542- Web search now takes today's date into context
543- Fixed a bug where stdio MCP servers were not terminating properly on exit
544
545## 1.0.35
546
547- Added support for MCP OAuth Authorization Server discovery
548
549## 1.0.34
550
551- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
552
553## 1.0.33
554
555- Improved logging functionality with session ID support
556- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
557- Improvements to plan mode
558
559## 1.0.32
560
561- Updated loopback config for litellm
562- Added forceLoginMethod setting to bypass login selection screen
563
564## 1.0.31
565
566- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
567
568## 1.0.30
569
570- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
571- Improved file path autocomplete with filename matching
572- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
573- Enhanced jq regex support for complex filters with pipes and select
574
575## 1.0.29
576
577- Improved CJK character support in cursor navigation and rendering
578
579## 1.0.28
580
581- Slash commands: Fix selector display during history navigation
582- Resizes images before upload to prevent API size limit errors
583- Added XDG_CONFIG_HOME support to configuration directory
584- Performance optimizations for memory usage
585- New attributes (terminal.type, language) in OpenTelemetry logging
586
587## 1.0.27
588
589- Streamable HTTP MCP servers are now supported
590- Remote MCP servers (SSE and HTTP) now support OAuth
591- MCP resources can now be @-mentioned
592- /resume slash command to switch conversations within Claude Code
593
594## 1.0.25
595
596- Slash commands: moved "project" and "user" prefixes to descriptions
597- Slash commands: improved reliability for command discovery
598- Improved support for Ghostty
599- Improved web search reliability
600
601## 1.0.24
602
603- Improved /mcp output
604- Fixed a bug where settings arrays got overwritten instead of merged
605
606## 1.0.23
607
608- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
609- Released Python SDK: pip install claude-code-sdk to get started
610
611## 1.0.22
612
613- SDK: Renamed `total_cost` to `total_cost_usd`
614
615## 1.0.21
616
617- Improved editing of files with tab-based indentation
618- Fix for tool_use without matching tool_result errors
619- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
620
621## 1.0.18
622
623- Added --add-dir CLI argument for specifying additional working directories
624- Added streaming input support without require -p flag
625- Improved startup performance and session storage performance
626- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
627- Added detailed MCP server tools display (/mcp)
628- MCP authentication and permission improvements
629- Added auto-reconnection for MCP SSE connections on disconnect
630- Fixed issue where pasted content was lost when dialogs appeared
631
632## 1.0.17
633
634- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
635- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
636- MCP server list UI improvements
637- Update Claude Code process title to display "claude" instead of "node"
638
639## 1.0.11
640
641- Claude Code can now also be used with a Claude Pro subscription
642- Added /upgrade for smoother switching to Claude Max plans
643- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
644- Improved shell configuration error handling
645- Improved todo list handling during compaction
646
647## 1.0.10
648
649- Added markdown table support
650- Improved streaming performance
651
652## 1.0.8
653
654- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
655- Increased default otel interval from 1s -> 5s
656- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
657- Fixed a regression where search tools unnecessarily asked for permissions
658- Added support for triggering thinking non-English languages
659- Improved compacting UI
660
661## 1.0.7
662
663- Renamed /allowed-tools -> /permissions
664- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
665- Deprecated claude config commands in favor of editing settings.json
666- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
667- Improved error handling for /install-github-app
668- Bugfixes, UI polish, and tool reliability improvements
669
670## 1.0.6
671
672- Improved edit reliability for tab-indented files
673- Respect CLAUDE_CONFIG_DIR everywhere
674- Reduced unnecessary tool permission prompts
675- Added support for symlinks in @file typeahead
676- Bugfixes, UI polish, and tool reliability improvements
677
678## 1.0.4
679
680- Fixed a bug where MCP tool errors weren't being parsed correctly
681
682## 1.0.1
683
684- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
685- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
686- Updated documentation links and OAuth process descriptions
687
688## 1.0.0
689
690- Claude Code is now generally available
691- Introducing Sonnet 4 and Opus 4 models
692
693## 0.2.125
694
695- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
696- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
697
698## 0.2.117
699
700- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
701- Introduced settings.cleanupPeriodDays
702- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
703- Introduced --debug mode
704
705## 0.2.108
706
707- You can now send messages to Claude while it works to steer Claude in real-time
708- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
709- Fixed a bug where thinking was not working in -p mode
710- Fixed a regression in /cost reporting
711- Deprecated MCP wizard interface in favor of other MCP commands
712- Lots of other bugfixes and improvements
713
714## 0.2.107
715
716- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
717
718## 0.2.106
719
720- MCP SSE server configs can now specify custom headers
721- Fixed a bug where MCP permission prompt didn't always show correctly
722
723## 0.2.105
724
725- Claude can now search the web
726- Moved system & account status to /status
727- Added word movement keybindings for Vim
728- Improved latency for startup, todo tool, and file edits
729
730## 0.2.102
731
732- Improved thinking triggering reliability
733- Improved @mention reliability for images and folders
734- You can now paste multiple large chunks into one prompt
735
736## 0.2.100
737
738- Fixed a crash caused by a stack overflow error
739- Made db storage optional; missing db support disables --continue and --resume
740
741## 0.2.98
742
743- Fixed an issue where auto-compact was running twice
744
745## 0.2.96
746
747- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
748
749## 0.2.93
750
751- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
752- Claude now has access to a Todo list that helps it stay on track and be more organized
753
754## 0.2.82
755
756- Added support for --disallowedTools
757- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
758
759## 0.2.75
760
761- Hit Enter to queue up additional messages while Claude is working
762- Drag in or copy/paste image files directly into the prompt
763- @-mention files to directly add them to context
764- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
765- Improved performance for filename auto-complete
766
767## 0.2.74
768
769- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
770- Task tool can now perform writes and run bash commands
771
772## 0.2.72
773
774- Updated spinner to indicate tokens loaded and tool usage
775
776## 0.2.70
777
778- Network commands like curl are now available for Claude to use
779- Claude can now run multiple web queries in parallel
780- Pressing ESC once immediately interrupts Claude in Auto-accept mode
781
782## 0.2.69
783
784- Fixed UI glitches with improved Select component behavior
785- Enhanced terminal output display with better text truncation logic
786
787## 0.2.67
788
789- Shared project permission rules can be saved in .claude/settings.json
790
791## 0.2.66
792
793- Print mode (-p) now supports streaming output via --output-format=stream-json
794- Fixed issue where pasting could trigger memory or bash mode unexpectedly
795
796## 0.2.63
797
798- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
799
800## 0.2.61
801
802- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
803- Enhanced image detection for more reliable clipboard paste functionality
804- Fixed an issue where ESC key could crash the conversation history selector
805
806## 0.2.59
807
808- Copy+paste images directly into your prompt
809- Improved progress indicators for bash and fetch tools
810- Bugfixes for non-interactive mode (-p)
811
812## 0.2.54
813
814- Quickly add to Memory by starting your message with '#'
815- Press ctrl+r to see full output for long tool results
816- Added support for MCP SSE transport
817
818## 0.2.53
819
820- New web fetch tool lets Claude view URLs that you paste in
821- Fixed a bug with JPEG detection
822
823## 0.2.50
824
825- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
826
827## 0.2.49
828
829- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
830
831## 0.2.47
832
833- Press Tab to auto-complete file and folder names
834- Press Shift + Tab to toggle auto-accept for file edits
835- Automatic conversation compaction for infinite conversation length (toggle with /config)
836
837## 0.2.44
838
839- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
840
841## 0.2.41
842
843- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
844- MCP server startup no longer blocks the app from starting up
845
846## 0.2.37
847
848- New /release-notes command lets you view release notes at any time
849- `claude config add/remove` commands now accept multiple values separated by commas or spaces
850
851## 0.2.36
852
853- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
854- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
855
856## 0.2.34
857
858- Vim bindings for text input - enable with /vim or /config
859
860## 0.2.32
861
862- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
863- Fix for some PersistentShell issues
864
865## 0.2.31
866
867- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
868- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
869
870## 0.2.30
871
872- Added ANSI color theme for better terminal compatibility
873- Fixed issue where slash command arguments weren't being sent properly
874- (Mac-only) API keys are now stored in macOS Keychain
875
876## 0.2.26
877
878- New /approved-tools command for managing tool permissions
879- Word-level diff display for improved code readability
880- Fuzzy matching for slash commands
881
882## 0.2.21
883
884- Fuzzy matching for /commands