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 5dd91a9 · on Nov 15, 2025 History
943 lines · 36.1 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.42
4
5- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
6
7## 2.0.41
8
9- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
10- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
11- Fixed incorrect labeling of user settings vs project settings in command descriptions
12- Fixed crash when plugin command hooks timeout during execution
13- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
14- Fixed broken security documentation links in trust dialogs and onboarding
15- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
16- ctrl-r history search landing on a slash command no longer cancels the search
17- SDK: Support custom timeouts for hooks
18- Allow more safe git commands to run without approval
19- Plugins: Added support for sharing and installing output styles
20- Teleporting a session from web will automatically set the upstream branch
21
22## 2.0.37
23
24- Fixed how idleness is computed for notifications
25- Hooks: Added matcher values for Notification hook events
26- Output Styles: Added `keep-coding-instructions` option to frontmatter
27
28## 2.0.36
29
30- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
31- Fixed queued messages being incorrectly executed as bash commands
32- Fixed input being lost when typing while a queued message is processed
33
34## 2.0.35
35
36- Improve fuzzy search results when searching commands
37- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
38- Added `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` environment variable to automatically exit SDK mode after a specified idle duration, useful for automated workflows and scripts
39- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
40- Fixed messages returning null `stop_reason` and `stop_sequence` values
41- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
42
43## 2.0.34
44
45- VSCode Extension: Added setting to configure the initial permission mode for new conversations
46- Improved file path suggestion performance with native Rust-based fuzzy finder
47- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
48- Fixed memory crash when reading or writing large files (especially base64-encoded images)
49
50## 2.0.33
51
52- Native binary installs now launch quicker.
53- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
54- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
55
56## 2.0.32
57
58- Un-deprecate output styles based on community feedback
59- Added `companyAnnouncements` setting for displaying announcements on startup
60- Fixed hook progress messages not updating correctly during PostToolUse hook execution
61
62## 2.0.31
63
64- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
65- Vertex: add support for Web Search on supported models
66- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
67- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
68- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
69- Fixed plugin uninstall not removing plugins
70
71## 2.0.30
72
73- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
74- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
75- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
76- Added prompt-based stop hooks
77- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
78- Enabled SSE MCP servers on native build
79- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
80- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
81- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
82- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
83- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
84- Fixed bug that saved session permissions to local settings
85- Fixed MCP tools not being available to sub-agents
86- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
87- Fixed delay when navigating through typeahead suggestions with arrow keys
88- VSCode: Restored selection indicator in input footer showing current file or code selection status
89
90## 2.0.28
91
92- Plan mode: introduced new Plan subagent
93- Subagents: claude can now choose to resume subagents
94- Subagents: claude can dynamically choose the model used by its subagents
95- SDK: added --max-budget-usd flag
96- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
97- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
98- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
99- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
100- Various other bug fixes
101
102## 2.0.27
103
104- New UI for permission prompts
105- Added current branch filtering and search to session resume screen for easier navigation
106- Fixed directory @-mention causing "No assistant message found" error
107- VSCode Extension: Add config setting to include .gitignored files in file searches
108- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
109
110## 2.0.25
111
112- 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
113
114## 2.0.24
115
116- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
117- Claude Code Web: Support for Web -> CLI teleport
118- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
119- Bedrock: Display awsAuthRefresh output when auth is required
120
121## 2.0.22
122
123- Fixed content layout shift when scrolling through slash commands
124- IDE: Add toggle to enable/disable thinking.
125- Fix bug causing duplicate permission prompts with parallel tool calls
126- Add support for enterprise managed MCP allowlist and denylist
127
128## 2.0.21
129
130- Support MCP `structuredContent` field in tool responses
131- Added an interactive question tool
132- Claude will now ask you questions more often in plan mode
133- Added Haiku 4.5 as a model option for Pro users
134- Fixed an issue where queued commands don't have access to previous messages' output
135
136## 2.0.20
137
138- Added support for Claude Skills
139
140## 2.0.19
141
142- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
143- Fixed a bug where Haiku was unnecessarily called in print mode
144
145## 2.0.17
146
147- Added Haiku 4.5 to model selector!
148- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
149- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
150- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
151- OTEL: support HTTP_PROXY and HTTPS_PROXY
152- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
153
154## 2.0.15
155
156- Fixed bug with resuming where previously created files needed to be read again before writing
157- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
158
159## 2.0.14
160
161- Fix @-mentioning MCP servers to toggle them on/off
162- Improve permission checks for bash with inline env vars
163- Fix ultrathink + thinking toggle
164- Reduce unnecessary logins
165- Document --system-prompt
166- Several improvements to rendering
167- Plugins UI polish
168
169## 2.0.13
170
171- Fixed `/plugin` not working on native build
172
173## 2.0.12
174
175- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
176- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
177- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
178- `/plugin validate` command for validating plugin structure and configuration
179- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
180- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
181- Comprehensive error messages and diagnostics via `/doctor` command
182- Avoid flickering in `/model` selector
183- Improvements to `/help`
184- Avoid mentioning hooks in `/resume` summaries
185- Changes to the "verbose" setting in `/config` now persist across sessions
186
187## 2.0.11
188
189- Reduced system prompt size by 1.4k tokens
190- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
191- Fixed Opus fallback rate limit errors appearing incorrectly
192- Fixed /add-dir command selecting wrong default tab
193
194## 2.0.10
195
196- Rewrote terminal renderer for buttery smooth UI
197- Enable/disable MCP servers by @mentioning, or in /mcp
198- Added tab completion for shell commands in bash mode
199- PreToolUse hooks can now modify tool inputs
200- Press Ctrl-G to edit your prompt in your system's configured text editor
201- Fixes for bash permission checks with environment variables in the command
202
203## 2.0.9
204
205- Fix regression where bash backgrounding stopped working
206
207## 2.0.8
208
209- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
210- IDE: Add drag-and-drop support for files and folders in chat
211- /context: Fix counting for thinking blocks
212- Improve message rendering for users with light themes on dark terminals
213- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
214
215## 2.0.5
216
217- IDE: Fix IME unintended message submission with Enter and Tab
218- IDE: Add "Open in Terminal" link in login screen
219- Fix unhandled OAuth expiration 401 API errors
220- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
221
222## 2.0.1
223
224- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
225- Various bug fixes and presentation improvements
226
227## 2.0.0
228
229- New native VS Code extension
230- Fresh coat of paint throughout the whole app
231- /rewind a conversation to undo code changes
232- /usage command to see plan limits
233- Tab to toggle thinking (sticky across sessions)
234- Ctrl-R to search history
235- Unshipped claude config command
236- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
237- SDK: The Claude Code SDK is now the Claude Agent SDK
238- Add subagents dynamically with `--agents` flag
239
240## 1.0.126
241
242- Enable /context command for Bedrock and Vertex
243- Add mTLS support for HTTP-based OpenTelemetry exporters
244
245## 1.0.124
246
247- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
248- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
249- No longer inform Claude of the list of allowed tools when permission is denied
250- Fixed security vulnerability in Bash tool permission checks
251- Improved VSCode extension performance for large files
252
253## 1.0.123
254
255- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
256- Fixed thinking mode triggering on negation phrases like "don't think"
257- Fixed rendering performance degradation during token streaming
258- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
259- Enhanced BashTool environment snapshot logging
260- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
261- Migrated --debug logging to a file, to enable easy tailing & filtering
262
263## 1.0.120
264
265- Fix input lag during typing, especially noticeable with large prompts
266- Improved VSCode extension command registry and sessions dialog user experience
267- Enhanced sessions dialog responsiveness and visual feedback
268- Fixed IDE compatibility issue by removing worktree support check
269- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
270
271## 1.0.119
272
273- Fix Windows issue where process visually freezes on entering interactive mode
274- Support dynamic headers for MCP servers via headersHelper configuration
275- Fix thinking mode not working in headless sessions
276- Fix slash commands now properly update allowed tools instead of replacing them
277
278## 1.0.117
279
280- Add Ctrl-R history search to recall previous commands like bash/zsh
281- Fix input lag while typing, especially on Windows
282- Add sed command to auto-allowed commands in acceptEdits mode
283- Fix Windows PATH comparison to be case-insensitive for drive letters
284- Add permissions management hint to /add-dir output
285
286## 1.0.115
287
288- Improve thinking mode display with enhanced visual effects
289- Type /t to temporarily disable thinking mode in your prompt
290- Improve path validation for glob and grep tools
291- Show condensed output for post-tool hooks to reduce visual clutter
292- Fix visual feedback when loading state completes
293- Improve UI consistency for permission request dialogs
294
295## 1.0.113
296
297- Deprecated piped input in interactive mode
298- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
299
300## 1.0.112
301
302- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
303- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
304- Hooks: Added systemMessage support for SessionEnd hooks
305- Added `spinnerTipsEnabled` setting to disable spinner tips
306- IDE: Various improvements and bug fixes
307
308## 1.0.111
309
310- /model now validates provided model names
311- Fixed Bash tool crashes caused by malformed shell syntax parsing
312
313## 1.0.110
314
315- /terminal-setup command now supports WezTerm
316- MCP: OAuth tokens now proactively refresh before expiration
317- Fixed reliability issues with background Bash processes
318
319## 1.0.109
320
321- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
322
323## 1.0.106
324
325- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
326
327## 1.0.97
328
329- Settings: /doctor now validates permission rule syntax and suggests corrections
330
331## 1.0.94
332
333- Vertex: add support for global endpoints for supported models
334- /memory command now allows direct editing of all imported memory files
335- SDK: Add custom tools as callbacks
336- Added /todos command to list current todo items
337
338## 1.0.93
339
340- Windows: Add alt + v shortcut for pasting images from clipboard
341- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
342
343## 1.0.90
344
345- Settings file changes take effect immediately - no restart required
346
347## 1.0.88
348
349- Fixed issue causing "OAuth authentication is currently not supported"
350- Status line input now includes `exceeds_200k_tokens`
351- Fixed incorrect usage tracking in /cost.
352- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
353- Bedrock: Updated default Sonnet model to Sonnet 4
354
355## 1.0.86
356
357- Added /context to help users self-serve debug context issues
358- SDK: Added UUID support for all SDK messages
359- SDK: Added `--replay-user-messages` to replay user messages back to stdout
360
361## 1.0.85
362
363- Status line input now includes session cost info
364- Hooks: Introduced SessionEnd hook
365
366## 1.0.84
367
368- Fix tool_use/tool_result id mismatch error when network is unstable
369- Fix Claude sometimes ignoring real-time steering when wrapping up a task
370- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
371- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
372
373## 1.0.83
374
375- @-mention: Support files with spaces in path
376- New shimmering spinner
377
378## 1.0.82
379
380- SDK: Add request cancellation support
381- SDK: New additionalDirectories option to search custom paths, improved slash command processing
382- Settings: Validation prevents invalid fields in .claude/settings.json files
383- MCP: Improve tool name consistency
384- Bash: Fix crash when Claude tries to automatically read large files
385
386## 1.0.81
387
388- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
389- Agents: Fix custom agent loading when agent files are unparsable
390
391## 1.0.80
392
393- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
394
395## 1.0.77
396
397- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
398- SDK: Add session support and permission denial tracking
399- Fix token limit errors in conversation summarization
400- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
401
402## 1.0.73
403
404- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
405- MCP: Press Esc to cancel OAuth authentication flows
406- Bash: Improved command validation and reduced false security warnings
407- UI: Enhanced spinner animations and status line visual hierarchy
408- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
409
410## 1.0.72
411
412- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
413
414## 1.0.71
415
416- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
417- Customizable status line: add your terminal prompt to Claude Code with /statusline
418
419## 1.0.70
420
421- Performance: Optimized message rendering for better performance with large contexts
422- Windows: Fixed native file search, ripgrep, and subagent functionality
423- Added support for @-mentions in slash command arguments
424
425## 1.0.69
426
427- Upgraded Opus to version 4.1
428
429## 1.0.68
430
431- Fix incorrect model names being used for certain commands like `/pr-comments`
432- 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.
433- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
434- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
435- SDK: Added canUseTool callback support for tool confirmation
436- Added `disableAllHooks` setting
437- Improved file suggestions performance in large repos
438
439## 1.0.65
440
441- IDE: Fixed connection stability issues and error handling for diagnostics
442- Windows: Fixed shell environment setup for users without .bashrc files
443
444## 1.0.64
445
446- Agents: Added model customization support - you can now specify which model an agent should use
447- Agents: Fixed unintended access to the recursive agent tool
448- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
449- SDK: Fixed user input tracking across multi-turn conversations
450- Added hidden files to file search and @-mention suggestions
451
452## 1.0.63
453
454- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
455
456## 1.0.62
457
458- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
459- Hooks: Added SessionStart hook for new session initialization
460- /add-dir command now supports typeahead for directory paths
461- Improved network connectivity check reliability
462
463## 1.0.61
464
465- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
466- Settings: Added `--settings` flag to load settings from a JSON file
467- Settings: Fixed resolution of settings files paths that are symlinks
468- OTEL: Fixed reporting of wrong organization after authentication changes
469- Slash commands: Fixed permissions checking for allowed-tools with Bash
470- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
471- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
472- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
473
474## 1.0.60
475
476- You can now create custom subagents for specialized tasks! Run /agents to get started
477
478## 1.0.59
479
480- SDK: Added tool confirmation support with canUseTool callback
481- SDK: Allow specifying env for spawned process
482- Hooks: Exposed PermissionDecision to hooks (including "ask")
483- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
484- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
485
486## 1.0.58
487
488- Added support for reading PDFs
489- MCP: Improved server health status display in 'claude mcp list'
490- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
491
492## 1.0.57
493
494- Added support for specifying a model in slash commands
495- Improved permission messages to help Claude understand allowed tools
496- Fix: Remove trailing newlines from bash output in terminal wrapping
497
498## 1.0.56
499
500- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
501- Fixes for WSL IDE detection
502- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
503
504## 1.0.55
505
506- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
507- Windows: fixed Ctrl+Z crash
508- SDK: Added ability to capture error logging
509- Add --system-prompt-file option to override system prompt in print mode
510
511## 1.0.54
512
513- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
514- Custom slash commands: Added argument-hint to frontmatter
515- Windows: OAuth uses port 45454 and properly constructs browser URL
516- Windows: mode switching now uses alt + m, and plan mode renders properly
517- Shell: Switch to in-memory shell snapshot to fix file-related errors
518
519## 1.0.53
520
521- Updated @-mention file truncation from 100 lines to 2000 lines
522- 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).
523
524## 1.0.52
525
526- Added support for MCP server instructions
527
528## 1.0.51
529
530- Added support for native Windows (requires Git for Windows)
531- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
532- Settings: /doctor can now help you identify and fix invalid setting files
533- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
534- Increased auto-compact warning threshold from 60% to 80%
535- Fixed an issue with handling user directories with spaces for shell snapshots
536- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
537- Custom slash commands: Fixed user-level commands in subdirectories
538- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
539
540## 1.0.48
541
542- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
543- Added progress messages to Bash tool based on the last 5 lines of command output
544- Added expanding variables support for MCP server configuration
545- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
546- Improved IDE extension path handling when Claude Code runs in WSL
547- Hooks: Added a PreCompact hook
548- Vim mode: Added c, f/F, t/T
549
550## 1.0.45
551
552- Redesigned Search (Grep) tool with new tool input parameters and features
553- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
554- Fixed config file corruption issue by enforcing atomic writes
555- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
556- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
557- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
558
559## 1.0.44
560
561- New /export command lets you quickly export a conversation for sharing
562- MCP: resource_link tool results are now supported
563- MCP: tool annotations and tool titles now display in /mcp view
564- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
565
566## 1.0.43
567
568- Fixed a bug where the theme selector was saving excessively
569- Hooks: Added EPIPE system error handling
570
571## 1.0.42
572
573- Added tilde (`~`) expansion support to `/add-dir` command
574
575## 1.0.41
576
577- Hooks: Split Stop hook triggering into Stop and SubagentStop
578- Hooks: Enabled optional timeout configuration for each command
579- Hooks: Added "hook_event_name" to hook input
580- Fixed a bug where MCP tools would display twice in tool list
581- New tool parameters JSON for Bash tool in `tool_decision` event
582
583## 1.0.40
584
585- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
586
587## 1.0.39
588
589- New Active Time metric in OpenTelemetry logging
590
591## 1.0.38
592
593- 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
594
595## 1.0.37
596
597- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
598
599## 1.0.36
600
601- Web search now takes today's date into context
602- Fixed a bug where stdio MCP servers were not terminating properly on exit
603
604## 1.0.35
605
606- Added support for MCP OAuth Authorization Server discovery
607
608## 1.0.34
609
610- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
611
612## 1.0.33
613
614- Improved logging functionality with session ID support
615- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
616- Improvements to plan mode
617
618## 1.0.32
619
620- Updated loopback config for litellm
621- Added forceLoginMethod setting to bypass login selection screen
622
623## 1.0.31
624
625- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
626
627## 1.0.30
628
629- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
630- Improved file path autocomplete with filename matching
631- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
632- Enhanced jq regex support for complex filters with pipes and select
633
634## 1.0.29
635
636- Improved CJK character support in cursor navigation and rendering
637
638## 1.0.28
639
640- Slash commands: Fix selector display during history navigation
641- Resizes images before upload to prevent API size limit errors
642- Added XDG_CONFIG_HOME support to configuration directory
643- Performance optimizations for memory usage
644- New attributes (terminal.type, language) in OpenTelemetry logging
645
646## 1.0.27
647
648- Streamable HTTP MCP servers are now supported
649- Remote MCP servers (SSE and HTTP) now support OAuth
650- MCP resources can now be @-mentioned
651- /resume slash command to switch conversations within Claude Code
652
653## 1.0.25
654
655- Slash commands: moved "project" and "user" prefixes to descriptions
656- Slash commands: improved reliability for command discovery
657- Improved support for Ghostty
658- Improved web search reliability
659
660## 1.0.24
661
662- Improved /mcp output
663- Fixed a bug where settings arrays got overwritten instead of merged
664
665## 1.0.23
666
667- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
668- Released Python SDK: pip install claude-code-sdk to get started
669
670## 1.0.22
671
672- SDK: Renamed `total_cost` to `total_cost_usd`
673
674## 1.0.21
675
676- Improved editing of files with tab-based indentation
677- Fix for tool_use without matching tool_result errors
678- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
679
680## 1.0.18
681
682- Added --add-dir CLI argument for specifying additional working directories
683- Added streaming input support without require -p flag
684- Improved startup performance and session storage performance
685- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
686- Added detailed MCP server tools display (/mcp)
687- MCP authentication and permission improvements
688- Added auto-reconnection for MCP SSE connections on disconnect
689- Fixed issue where pasted content was lost when dialogs appeared
690
691## 1.0.17
692
693- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
694- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
695- MCP server list UI improvements
696- Update Claude Code process title to display "claude" instead of "node"
697
698## 1.0.11
699
700- Claude Code can now also be used with a Claude Pro subscription
701- Added /upgrade for smoother switching to Claude Max plans
702- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
703- Improved shell configuration error handling
704- Improved todo list handling during compaction
705
706## 1.0.10
707
708- Added markdown table support
709- Improved streaming performance
710
711## 1.0.8
712
713- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
714- Increased default otel interval from 1s -> 5s
715- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
716- Fixed a regression where search tools unnecessarily asked for permissions
717- Added support for triggering thinking non-English languages
718- Improved compacting UI
719
720## 1.0.7
721
722- Renamed /allowed-tools -> /permissions
723- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
724- Deprecated claude config commands in favor of editing settings.json
725- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
726- Improved error handling for /install-github-app
727- Bugfixes, UI polish, and tool reliability improvements
728
729## 1.0.6
730
731- Improved edit reliability for tab-indented files
732- Respect CLAUDE_CONFIG_DIR everywhere
733- Reduced unnecessary tool permission prompts
734- Added support for symlinks in @file typeahead
735- Bugfixes, UI polish, and tool reliability improvements
736
737## 1.0.4
738
739- Fixed a bug where MCP tool errors weren't being parsed correctly
740
741## 1.0.1
742
743- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
744- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
745- Updated documentation links and OAuth process descriptions
746
747## 1.0.0
748
749- Claude Code is now generally available
750- Introducing Sonnet 4 and Opus 4 models
751
752## 0.2.125
753
754- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
755- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
756
757## 0.2.117
758
759- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
760- Introduced settings.cleanupPeriodDays
761- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
762- Introduced --debug mode
763
764## 0.2.108
765
766- You can now send messages to Claude while it works to steer Claude in real-time
767- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
768- Fixed a bug where thinking was not working in -p mode
769- Fixed a regression in /cost reporting
770- Deprecated MCP wizard interface in favor of other MCP commands
771- Lots of other bugfixes and improvements
772
773## 0.2.107
774
775- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
776
777## 0.2.106
778
779- MCP SSE server configs can now specify custom headers
780- Fixed a bug where MCP permission prompt didn't always show correctly
781
782## 0.2.105
783
784- Claude can now search the web
785- Moved system & account status to /status
786- Added word movement keybindings for Vim
787- Improved latency for startup, todo tool, and file edits
788
789## 0.2.102
790
791- Improved thinking triggering reliability
792- Improved @mention reliability for images and folders
793- You can now paste multiple large chunks into one prompt
794
795## 0.2.100
796
797- Fixed a crash caused by a stack overflow error
798- Made db storage optional; missing db support disables --continue and --resume
799
800## 0.2.98
801
802- Fixed an issue where auto-compact was running twice
803
804## 0.2.96
805
806- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
807
808## 0.2.93
809
810- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
811- Claude now has access to a Todo list that helps it stay on track and be more organized
812
813## 0.2.82
814
815- Added support for --disallowedTools
816- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
817
818## 0.2.75
819
820- Hit Enter to queue up additional messages while Claude is working
821- Drag in or copy/paste image files directly into the prompt
822- @-mention files to directly add them to context
823- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
824- Improved performance for filename auto-complete
825
826## 0.2.74
827
828- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
829- Task tool can now perform writes and run bash commands
830
831## 0.2.72
832
833- Updated spinner to indicate tokens loaded and tool usage
834
835## 0.2.70
836
837- Network commands like curl are now available for Claude to use
838- Claude can now run multiple web queries in parallel
839- Pressing ESC once immediately interrupts Claude in Auto-accept mode
840
841## 0.2.69
842
843- Fixed UI glitches with improved Select component behavior
844- Enhanced terminal output display with better text truncation logic
845
846## 0.2.67
847
848- Shared project permission rules can be saved in .claude/settings.json
849
850## 0.2.66
851
852- Print mode (-p) now supports streaming output via --output-format=stream-json
853- Fixed issue where pasting could trigger memory or bash mode unexpectedly
854
855## 0.2.63
856
857- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
858
859## 0.2.61
860
861- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
862- Enhanced image detection for more reliable clipboard paste functionality
863- Fixed an issue where ESC key could crash the conversation history selector
864
865## 0.2.59
866
867- Copy+paste images directly into your prompt
868- Improved progress indicators for bash and fetch tools
869- Bugfixes for non-interactive mode (-p)
870
871## 0.2.54
872
873- Quickly add to Memory by starting your message with '#'
874- Press ctrl+r to see full output for long tool results
875- Added support for MCP SSE transport
876
877## 0.2.53
878
879- New web fetch tool lets Claude view URLs that you paste in
880- Fixed a bug with JPEG detection
881
882## 0.2.50
883
884- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
885
886## 0.2.49
887
888- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
889
890## 0.2.47
891
892- Press Tab to auto-complete file and folder names
893- Press Shift + Tab to toggle auto-accept for file edits
894- Automatic conversation compaction for infinite conversation length (toggle with /config)
895
896## 0.2.44
897
898- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
899
900## 0.2.41
901
902- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
903- MCP server startup no longer blocks the app from starting up
904
905## 0.2.37
906
907- New /release-notes command lets you view release notes at any time
908- `claude config add/remove` commands now accept multiple values separated by commas or spaces
909
910## 0.2.36
911
912- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
913- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
914
915## 0.2.34
916
917- Vim bindings for text input - enable with /vim or /config
918
919## 0.2.32
920
921- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
922- Fix for some PersistentShell issues
923
924## 0.2.31
925
926- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
927- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
928
929## 0.2.30
930
931- Added ANSI color theme for better terminal compatibility
932- Fixed issue where slash command arguments weren't being sent properly
933- (Mac-only) API keys are now stored in macOS Keychain
934
935## 0.2.26
936
937- New /approved-tools command for managing tool permissions
938- Word-level diff display for improved code readability
939- Fuzzy matching for slash commands
940
941## 0.2.21
942
943- Fuzzy matching for /commands