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 2192c86 · on Dec 12, 2025 History
1,100 lines · 44.3 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.67
4
5- Claude now suggests prompts to speed up your workflow: press Tab to accept or Enter to submit
6- Thinking mode is now enabled by default for Opus 4.5
7- Thinking mode configuration has moved to /config
8- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
9- Show reason why autoupdater is disabled in `/doctor`
10- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
11- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
12- Fixed scroll position resetting after deleting a permission rule in `/permissions`
13- Fixed word deletion (opt+delete) and word navigation (opt+arrow) not working correctly with non-Latin text such as Cyrillic, Greek, Arabic, Hebrew, Thai, and Chinese
14- Fixed `claude install --force` not bypassing stale lock files
15- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
16- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
17
18## 2.0.65
19
20- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
21- Added context window information to status line input
22- Added `fileSuggestion` setting for custom `@` file search commands
23- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
24- Fixed prompt not being saved to history when aborting a query with Escape
25- Fixed Read tool image handling to identify format from bytes instead of file extension
26
27## 2.0.64
28
29- Made auto-compacting instant
30- Agents and bash commands can run asynchronously and send messages to wake up the main agent
31- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
32- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
33- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
34- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
35- Fixed auto-loading .env when using native installer
36- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
37- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
38- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
39- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
40- Bedrock: Improve efficiency of token counting
41- Bedrock: Add support for `aws login` AWS Management Console credentials
42- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
43
44## 2.0.62
45
46- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
47- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
48- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
49- Fixed slash command selection not working when multiple commands share the same name
50- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
51- Fixed running versions getting removed because lock file incorrectly going stale
52- Fixed IDE diff tab not closing when rejecting file changes
53
54## 2.0.61
55
56- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
57
58## 2.0.60
59
60- Added background agent support. Agents run in the background while you work
61- Added --disable-slash-commands CLI flag to disable all slash commands
62- Added model name to "Co-Authored-By" commit messages
63- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
64- Updated Fetch to skip summarization for pre-approved websites
65- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
66
67## 2.0.59
68
69- Added --agent CLI flag to override the agent setting for the current session
70- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
71- VS Code: Fixed .claude.json config file being read from incorrect location
72
73## 2.0.58
74
75- Pro users now have access to Opus 4.5 as part of their subscription!
76- Fixed timer duration showing "11m 60s" instead of "12m 0s"
77- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
78
79## 2.0.57
80
81- Added feedback input when rejecting plans, allowing users to tell Claude what to change
82- VSCode: Added streaming message support for real-time response display
83
84## 2.0.56
85
86- Added setting to enable/disable terminal progress bar (OSC 9;4)
87- VSCode Extension: Added support for VS Code's secondary sidebar (VS Code 1.97+), allowing Claude Code to be displayed in the right sidebar while keeping the file explorer on the left. Requires setting sidebar as Preferred Location in the config.
88
89## 2.0.55
90
91- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
92- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
93- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
94- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
95
96## 2.0.54
97
98- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
99- Fix issue with excessive iTerm notifications
100
101## 2.0.52
102
103- Fixed duplicate message display when starting Claude with a command line argument
104- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
105- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
106- Permit some uses of `$!` in bash commands
107
108## 2.0.51
109
110- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
111- Introducing Claude Code for Desktop: https://claude.com/download
112- To give you room to try out our new model, we've updated usage limits for Claude Code users. See the Claude Opus 4.5 blog for full details
113- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
114- Plan Mode now builds more precise plans and executes more thoroughly
115- Usage limit notifications now easier to understand
116- Switched `/usage` back to "% used"
117- Fixed handling of thinking errors
118- Fixed performance regression
119
120## 2.0.50
121
122- Fixed bug preventing calling MCP tools that have nested references in their input schemas
123- Silenced a noisy but harmless error during upgrades
124- Improved ultrathink text display
125- Improved clarity of 5-hour session limit warning message
126
127## 2.0.49
128
129- Added readline-style ctrl-y for pasting deleted text
130- Improved clarity of usage limit warning message
131- Fixed handling of subagent permissions
132
133## 2.0.47
134
135- Improved error messages and validation for `claude --teleport`
136- Improved error handling in `/usage`
137- Fixed race condition with history entry not getting logged at exit
138- Fixed Vertex AI configuration not being applied from `settings.json`
139
140## 2.0.46
141
142- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
143
144## 2.0.45
145
146- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
147- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
148- Send background tasks to Claude Code on the web by starting a message with `&`
149
150## 2.0.43
151
152- Added `permissionMode` field for custom agents
153- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
154- Added skills frontmatter field to declare skills to auto-load for subagents
155- Added the `SubagentStart` hook event
156- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
157- Fixed duplicate rendering of some messages in the UI
158- Fixed some visual flickers
159- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
160
161## 2.0.42
162
163- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
164
165## 2.0.41
166
167- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
168- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
169- Fixed incorrect labeling of user settings vs project settings in command descriptions
170- Fixed crash when plugin command hooks timeout during execution
171- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
172- Fixed broken security documentation links in trust dialogs and onboarding
173- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
174- ctrl-r history search landing on a slash command no longer cancels the search
175- SDK: Support custom timeouts for hooks
176- Allow more safe git commands to run without approval
177- Plugins: Added support for sharing and installing output styles
178- Teleporting a session from web will automatically set the upstream branch
179
180## 2.0.37
181
182- Fixed how idleness is computed for notifications
183- Hooks: Added matcher values for Notification hook events
184- Output Styles: Added `keep-coding-instructions` option to frontmatter
185
186## 2.0.36
187
188- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
189- Fixed queued messages being incorrectly executed as bash commands
190- Fixed input being lost when typing while a queued message is processed
191
192## 2.0.35
193
194- Improve fuzzy search results when searching commands
195- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
196- 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
197- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
198- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
199
200## 2.0.34
201
202- VSCode Extension: Added setting to configure the initial permission mode for new conversations
203- Improved file path suggestion performance with native Rust-based fuzzy finder
204- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
205- Fixed memory crash when reading or writing large files (especially base64-encoded images)
206
207## 2.0.33
208
209- Native binary installs now launch quicker.
210- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
211- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
212
213## 2.0.32
214
215- Un-deprecate output styles based on community feedback
216- Added `companyAnnouncements` setting for displaying announcements on startup
217- Fixed hook progress messages not updating correctly during PostToolUse hook execution
218
219## 2.0.31
220
221- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
222- Vertex: add support for Web Search on supported models
223- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
224- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
225- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
226- Fixed plugin uninstall not removing plugins
227
228## 2.0.30
229
230- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
231- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
232- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
233- Added prompt-based stop hooks
234- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
235- Enabled SSE MCP servers on native build
236- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
237- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
238- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
239- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
240- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
241- Fixed bug that saved session permissions to local settings
242- Fixed MCP tools not being available to sub-agents
243- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
244- Fixed delay when navigating through typeahead suggestions with arrow keys
245- VSCode: Restored selection indicator in input footer showing current file or code selection status
246
247## 2.0.28
248
249- Plan mode: introduced new Plan subagent
250- Subagents: claude can now choose to resume subagents
251- Subagents: claude can dynamically choose the model used by its subagents
252- SDK: added --max-budget-usd flag
253- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
254- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
255- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
256- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
257- Various other bug fixes
258
259## 2.0.27
260
261- New UI for permission prompts
262- Added current branch filtering and search to session resume screen for easier navigation
263- Fixed directory @-mention causing "No assistant message found" error
264- VSCode Extension: Add config setting to include .gitignored files in file searches
265- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
266
267## 2.0.25
268
269- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://platform.claude.com/docs/en/agent-sdk/migration-guide
270
271## 2.0.24
272
273- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
274- Claude Code Web: Support for Web -> CLI teleport
275- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
276- Bedrock: Display awsAuthRefresh output when auth is required
277
278## 2.0.22
279
280- Fixed content layout shift when scrolling through slash commands
281- IDE: Add toggle to enable/disable thinking.
282- Fix bug causing duplicate permission prompts with parallel tool calls
283- Add support for enterprise managed MCP allowlist and denylist
284
285## 2.0.21
286
287- Support MCP `structuredContent` field in tool responses
288- Added an interactive question tool
289- Claude will now ask you questions more often in plan mode
290- Added Haiku 4.5 as a model option for Pro users
291- Fixed an issue where queued commands don't have access to previous messages' output
292
293## 2.0.20
294
295- Added support for Claude Skills
296
297## 2.0.19
298
299- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
300- Fixed a bug where Haiku was unnecessarily called in print mode
301
302## 2.0.17
303
304- Added Haiku 4.5 to model selector!
305- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
306- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
307- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
308- OTEL: support HTTP_PROXY and HTTPS_PROXY
309- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
310
311## 2.0.15
312
313- Fixed bug with resuming where previously created files needed to be read again before writing
314- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
315
316## 2.0.14
317
318- Fix @-mentioning MCP servers to toggle them on/off
319- Improve permission checks for bash with inline env vars
320- Fix ultrathink + thinking toggle
321- Reduce unnecessary logins
322- Document --system-prompt
323- Several improvements to rendering
324- Plugins UI polish
325
326## 2.0.13
327
328- Fixed `/plugin` not working on native build
329
330## 2.0.12
331
332- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
333- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
334- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
335- `/plugin validate` command for validating plugin structure and configuration
336- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
337- Plugin documentation available at https://code.claude.com/docs/en/plugins
338- Comprehensive error messages and diagnostics via `/doctor` command
339- Avoid flickering in `/model` selector
340- Improvements to `/help`
341- Avoid mentioning hooks in `/resume` summaries
342- Changes to the "verbose" setting in `/config` now persist across sessions
343
344## 2.0.11
345
346- Reduced system prompt size by 1.4k tokens
347- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
348- Fixed Opus fallback rate limit errors appearing incorrectly
349- Fixed /add-dir command selecting wrong default tab
350
351## 2.0.10
352
353- Rewrote terminal renderer for buttery smooth UI
354- Enable/disable MCP servers by @mentioning, or in /mcp
355- Added tab completion for shell commands in bash mode
356- PreToolUse hooks can now modify tool inputs
357- Press Ctrl-G to edit your prompt in your system's configured text editor
358- Fixes for bash permission checks with environment variables in the command
359
360## 2.0.9
361
362- Fix regression where bash backgrounding stopped working
363
364## 2.0.8
365
366- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
367- IDE: Add drag-and-drop support for files and folders in chat
368- /context: Fix counting for thinking blocks
369- Improve message rendering for users with light themes on dark terminals
370- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
371
372## 2.0.5
373
374- IDE: Fix IME unintended message submission with Enter and Tab
375- IDE: Add "Open in Terminal" link in login screen
376- Fix unhandled OAuth expiration 401 API errors
377- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
378
379## 2.0.1
380
381- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
382- Various bug fixes and presentation improvements
383
384## 2.0.0
385
386- New native VS Code extension
387- Fresh coat of paint throughout the whole app
388- /rewind a conversation to undo code changes
389- /usage command to see plan limits
390- Tab to toggle thinking (sticky across sessions)
391- Ctrl-R to search history
392- Unshipped claude config command
393- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
394- SDK: The Claude Code SDK is now the Claude Agent SDK
395- Add subagents dynamically with `--agents` flag
396
397## 1.0.126
398
399- Enable /context command for Bedrock and Vertex
400- Add mTLS support for HTTP-based OpenTelemetry exporters
401
402## 1.0.124
403
404- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
405- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
406- No longer inform Claude of the list of allowed tools when permission is denied
407- Fixed security vulnerability in Bash tool permission checks
408- Improved VSCode extension performance for large files
409
410## 1.0.123
411
412- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
413- Fixed thinking mode triggering on negation phrases like "don't think"
414- Fixed rendering performance degradation during token streaming
415- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
416- Enhanced BashTool environment snapshot logging
417- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
418- Migrated --debug logging to a file, to enable easy tailing & filtering
419
420## 1.0.120
421
422- Fix input lag during typing, especially noticeable with large prompts
423- Improved VSCode extension command registry and sessions dialog user experience
424- Enhanced sessions dialog responsiveness and visual feedback
425- Fixed IDE compatibility issue by removing worktree support check
426- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
427
428## 1.0.119
429
430- Fix Windows issue where process visually freezes on entering interactive mode
431- Support dynamic headers for MCP servers via headersHelper configuration
432- Fix thinking mode not working in headless sessions
433- Fix slash commands now properly update allowed tools instead of replacing them
434
435## 1.0.117
436
437- Add Ctrl-R history search to recall previous commands like bash/zsh
438- Fix input lag while typing, especially on Windows
439- Add sed command to auto-allowed commands in acceptEdits mode
440- Fix Windows PATH comparison to be case-insensitive for drive letters
441- Add permissions management hint to /add-dir output
442
443## 1.0.115
444
445- Improve thinking mode display with enhanced visual effects
446- Type /t to temporarily disable thinking mode in your prompt
447- Improve path validation for glob and grep tools
448- Show condensed output for post-tool hooks to reduce visual clutter
449- Fix visual feedback when loading state completes
450- Improve UI consistency for permission request dialogs
451
452## 1.0.113
453
454- Deprecated piped input in interactive mode
455- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
456
457## 1.0.112
458
459- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
460- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
461- Hooks: Added systemMessage support for SessionEnd hooks
462- Added `spinnerTipsEnabled` setting to disable spinner tips
463- IDE: Various improvements and bug fixes
464
465## 1.0.111
466
467- /model now validates provided model names
468- Fixed Bash tool crashes caused by malformed shell syntax parsing
469
470## 1.0.110
471
472- /terminal-setup command now supports WezTerm
473- MCP: OAuth tokens now proactively refresh before expiration
474- Fixed reliability issues with background Bash processes
475
476## 1.0.109
477
478- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
479
480## 1.0.106
481
482- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
483
484## 1.0.97
485
486- Settings: /doctor now validates permission rule syntax and suggests corrections
487
488## 1.0.94
489
490- Vertex: add support for global endpoints for supported models
491- /memory command now allows direct editing of all imported memory files
492- SDK: Add custom tools as callbacks
493- Added /todos command to list current todo items
494
495## 1.0.93
496
497- Windows: Add alt + v shortcut for pasting images from clipboard
498- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
499
500## 1.0.90
501
502- Settings file changes take effect immediately - no restart required
503
504## 1.0.88
505
506- Fixed issue causing "OAuth authentication is currently not supported"
507- Status line input now includes `exceeds_200k_tokens`
508- Fixed incorrect usage tracking in /cost.
509- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
510- Bedrock: Updated default Sonnet model to Sonnet 4
511
512## 1.0.86
513
514- Added /context to help users self-serve debug context issues
515- SDK: Added UUID support for all SDK messages
516- SDK: Added `--replay-user-messages` to replay user messages back to stdout
517
518## 1.0.85
519
520- Status line input now includes session cost info
521- Hooks: Introduced SessionEnd hook
522
523## 1.0.84
524
525- Fix tool_use/tool_result id mismatch error when network is unstable
526- Fix Claude sometimes ignoring real-time steering when wrapping up a task
527- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
528- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
529
530## 1.0.83
531
532- @-mention: Support files with spaces in path
533- New shimmering spinner
534
535## 1.0.82
536
537- SDK: Add request cancellation support
538- SDK: New additionalDirectories option to search custom paths, improved slash command processing
539- Settings: Validation prevents invalid fields in .claude/settings.json files
540- MCP: Improve tool name consistency
541- Bash: Fix crash when Claude tries to automatically read large files
542
543## 1.0.81
544
545- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
546- Agents: Fix custom agent loading when agent files are unparsable
547
548## 1.0.80
549
550- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
551
552## 1.0.77
553
554- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
555- SDK: Add session support and permission denial tracking
556- Fix token limit errors in conversation summarization
557- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
558
559## 1.0.73
560
561- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
562- MCP: Press Esc to cancel OAuth authentication flows
563- Bash: Improved command validation and reduced false security warnings
564- UI: Enhanced spinner animations and status line visual hierarchy
565- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
566
567## 1.0.72
568
569- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
570
571## 1.0.71
572
573- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
574- Customizable status line: add your terminal prompt to Claude Code with /statusline
575
576## 1.0.70
577
578- Performance: Optimized message rendering for better performance with large contexts
579- Windows: Fixed native file search, ripgrep, and subagent functionality
580- Added support for @-mentions in slash command arguments
581
582## 1.0.69
583
584- Upgraded Opus to version 4.1
585
586## 1.0.68
587
588- Fix incorrect model names being used for certain commands like `/pr-comments`
589- 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.
590- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
591- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
592- SDK: Added canUseTool callback support for tool confirmation
593- Added `disableAllHooks` setting
594- Improved file suggestions performance in large repos
595
596## 1.0.65
597
598- IDE: Fixed connection stability issues and error handling for diagnostics
599- Windows: Fixed shell environment setup for users without .bashrc files
600
601## 1.0.64
602
603- Agents: Added model customization support - you can now specify which model an agent should use
604- Agents: Fixed unintended access to the recursive agent tool
605- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
606- SDK: Fixed user input tracking across multi-turn conversations
607- Added hidden files to file search and @-mention suggestions
608
609## 1.0.63
610
611- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
612
613## 1.0.62
614
615- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
616- Hooks: Added SessionStart hook for new session initialization
617- /add-dir command now supports typeahead for directory paths
618- Improved network connectivity check reliability
619
620## 1.0.61
621
622- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
623- Settings: Added `--settings` flag to load settings from a JSON file
624- Settings: Fixed resolution of settings files paths that are symlinks
625- OTEL: Fixed reporting of wrong organization after authentication changes
626- Slash commands: Fixed permissions checking for allowed-tools with Bash
627- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
628- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
629- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
630
631## 1.0.60
632
633- You can now create custom subagents for specialized tasks! Run /agents to get started
634
635## 1.0.59
636
637- SDK: Added tool confirmation support with canUseTool callback
638- SDK: Allow specifying env for spawned process
639- Hooks: Exposed PermissionDecision to hooks (including "ask")
640- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
641- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
642
643## 1.0.58
644
645- Added support for reading PDFs
646- MCP: Improved server health status display in 'claude mcp list'
647- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
648
649## 1.0.57
650
651- Added support for specifying a model in slash commands
652- Improved permission messages to help Claude understand allowed tools
653- Fix: Remove trailing newlines from bash output in terminal wrapping
654
655## 1.0.56
656
657- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
658- Fixes for WSL IDE detection
659- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
660
661## 1.0.55
662
663- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
664- Windows: fixed Ctrl+Z crash
665- SDK: Added ability to capture error logging
666- Add --system-prompt-file option to override system prompt in print mode
667
668## 1.0.54
669
670- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
671- Custom slash commands: Added argument-hint to frontmatter
672- Windows: OAuth uses port 45454 and properly constructs browser URL
673- Windows: mode switching now uses alt + m, and plan mode renders properly
674- Shell: Switch to in-memory shell snapshot to fix file-related errors
675
676## 1.0.53
677
678- Updated @-mention file truncation from 100 lines to 2000 lines
679- 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).
680
681## 1.0.52
682
683- Added support for MCP server instructions
684
685## 1.0.51
686
687- Added support for native Windows (requires Git for Windows)
688- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
689- Settings: /doctor can now help you identify and fix invalid setting files
690- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
691- Increased auto-compact warning threshold from 60% to 80%
692- Fixed an issue with handling user directories with spaces for shell snapshots
693- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
694- Custom slash commands: Fixed user-level commands in subdirectories
695- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
696
697## 1.0.48
698
699- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
700- Added progress messages to Bash tool based on the last 5 lines of command output
701- Added expanding variables support for MCP server configuration
702- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
703- Improved IDE extension path handling when Claude Code runs in WSL
704- Hooks: Added a PreCompact hook
705- Vim mode: Added c, f/F, t/T
706
707## 1.0.45
708
709- Redesigned Search (Grep) tool with new tool input parameters and features
710- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
711- Fixed config file corruption issue by enforcing atomic writes
712- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
713- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
714- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
715
716## 1.0.44
717
718- New /export command lets you quickly export a conversation for sharing
719- MCP: resource_link tool results are now supported
720- MCP: tool annotations and tool titles now display in /mcp view
721- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
722
723## 1.0.43
724
725- Fixed a bug where the theme selector was saving excessively
726- Hooks: Added EPIPE system error handling
727
728## 1.0.42
729
730- Added tilde (`~`) expansion support to `/add-dir` command
731
732## 1.0.41
733
734- Hooks: Split Stop hook triggering into Stop and SubagentStop
735- Hooks: Enabled optional timeout configuration for each command
736- Hooks: Added "hook_event_name" to hook input
737- Fixed a bug where MCP tools would display twice in tool list
738- New tool parameters JSON for Bash tool in `tool_decision` event
739
740## 1.0.40
741
742- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
743
744## 1.0.39
745
746- New Active Time metric in OpenTelemetry logging
747
748## 1.0.38
749
750- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
751
752## 1.0.37
753
754- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
755
756## 1.0.36
757
758- Web search now takes today's date into context
759- Fixed a bug where stdio MCP servers were not terminating properly on exit
760
761## 1.0.35
762
763- Added support for MCP OAuth Authorization Server discovery
764
765## 1.0.34
766
767- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
768
769## 1.0.33
770
771- Improved logging functionality with session ID support
772- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
773- Improvements to plan mode
774
775## 1.0.32
776
777- Updated loopback config for litellm
778- Added forceLoginMethod setting to bypass login selection screen
779
780## 1.0.31
781
782- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
783
784## 1.0.30
785
786- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
787- Improved file path autocomplete with filename matching
788- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
789- Enhanced jq regex support for complex filters with pipes and select
790
791## 1.0.29
792
793- Improved CJK character support in cursor navigation and rendering
794
795## 1.0.28
796
797- Slash commands: Fix selector display during history navigation
798- Resizes images before upload to prevent API size limit errors
799- Added XDG_CONFIG_HOME support to configuration directory
800- Performance optimizations for memory usage
801- New attributes (terminal.type, language) in OpenTelemetry logging
802
803## 1.0.27
804
805- Streamable HTTP MCP servers are now supported
806- Remote MCP servers (SSE and HTTP) now support OAuth
807- MCP resources can now be @-mentioned
808- /resume slash command to switch conversations within Claude Code
809
810## 1.0.25
811
812- Slash commands: moved "project" and "user" prefixes to descriptions
813- Slash commands: improved reliability for command discovery
814- Improved support for Ghostty
815- Improved web search reliability
816
817## 1.0.24
818
819- Improved /mcp output
820- Fixed a bug where settings arrays got overwritten instead of merged
821
822## 1.0.23
823
824- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
825- Released Python SDK: pip install claude-code-sdk to get started
826
827## 1.0.22
828
829- SDK: Renamed `total_cost` to `total_cost_usd`
830
831## 1.0.21
832
833- Improved editing of files with tab-based indentation
834- Fix for tool_use without matching tool_result errors
835- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
836
837## 1.0.18
838
839- Added --add-dir CLI argument for specifying additional working directories
840- Added streaming input support without require -p flag
841- Improved startup performance and session storage performance
842- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
843- Added detailed MCP server tools display (/mcp)
844- MCP authentication and permission improvements
845- Added auto-reconnection for MCP SSE connections on disconnect
846- Fixed issue where pasted content was lost when dialogs appeared
847
848## 1.0.17
849
850- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
851- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
852- MCP server list UI improvements
853- Update Claude Code process title to display "claude" instead of "node"
854
855## 1.0.11
856
857- Claude Code can now also be used with a Claude Pro subscription
858- Added /upgrade for smoother switching to Claude Max plans
859- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
860- Improved shell configuration error handling
861- Improved todo list handling during compaction
862
863## 1.0.10
864
865- Added markdown table support
866- Improved streaming performance
867
868## 1.0.8
869
870- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
871- Increased default otel interval from 1s -> 5s
872- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
873- Fixed a regression where search tools unnecessarily asked for permissions
874- Added support for triggering thinking non-English languages
875- Improved compacting UI
876
877## 1.0.7
878
879- Renamed /allowed-tools -> /permissions
880- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
881- Deprecated claude config commands in favor of editing settings.json
882- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
883- Improved error handling for /install-github-app
884- Bugfixes, UI polish, and tool reliability improvements
885
886## 1.0.6
887
888- Improved edit reliability for tab-indented files
889- Respect CLAUDE_CONFIG_DIR everywhere
890- Reduced unnecessary tool permission prompts
891- Added support for symlinks in @file typeahead
892- Bugfixes, UI polish, and tool reliability improvements
893
894## 1.0.4
895
896- Fixed a bug where MCP tool errors weren't being parsed correctly
897
898## 1.0.1
899
900- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
901- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
902- Updated documentation links and OAuth process descriptions
903
904## 1.0.0
905
906- Claude Code is now generally available
907- Introducing Sonnet 4 and Opus 4 models
908
909## 0.2.125
910
911- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
912- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
913
914## 0.2.117
915
916- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
917- Introduced settings.cleanupPeriodDays
918- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
919- Introduced --debug mode
920
921## 0.2.108
922
923- You can now send messages to Claude while it works to steer Claude in real-time
924- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
925- Fixed a bug where thinking was not working in -p mode
926- Fixed a regression in /cost reporting
927- Deprecated MCP wizard interface in favor of other MCP commands
928- Lots of other bugfixes and improvements
929
930## 0.2.107
931
932- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
933
934## 0.2.106
935
936- MCP SSE server configs can now specify custom headers
937- Fixed a bug where MCP permission prompt didn't always show correctly
938
939## 0.2.105
940
941- Claude can now search the web
942- Moved system & account status to /status
943- Added word movement keybindings for Vim
944- Improved latency for startup, todo tool, and file edits
945
946## 0.2.102
947
948- Improved thinking triggering reliability
949- Improved @mention reliability for images and folders
950- You can now paste multiple large chunks into one prompt
951
952## 0.2.100
953
954- Fixed a crash caused by a stack overflow error
955- Made db storage optional; missing db support disables --continue and --resume
956
957## 0.2.98
958
959- Fixed an issue where auto-compact was running twice
960
961## 0.2.96
962
963- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
964
965## 0.2.93
966
967- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
968- Claude now has access to a Todo list that helps it stay on track and be more organized
969
970## 0.2.82
971
972- Added support for --disallowedTools
973- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
974
975## 0.2.75
976
977- Hit Enter to queue up additional messages while Claude is working
978- Drag in or copy/paste image files directly into the prompt
979- @-mention files to directly add them to context
980- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
981- Improved performance for filename auto-complete
982
983## 0.2.74
984
985- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
986- Task tool can now perform writes and run bash commands
987
988## 0.2.72
989
990- Updated spinner to indicate tokens loaded and tool usage
991
992## 0.2.70
993
994- Network commands like curl are now available for Claude to use
995- Claude can now run multiple web queries in parallel
996- Pressing ESC once immediately interrupts Claude in Auto-accept mode
997
998## 0.2.69
999
1000- Fixed UI glitches with improved Select component behavior
1001- Enhanced terminal output display with better text truncation logic
1002
1003## 0.2.67
1004
1005- Shared project permission rules can be saved in .claude/settings.json
1006
1007## 0.2.66
1008
1009- Print mode (-p) now supports streaming output via --output-format=stream-json
1010- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1011
1012## 0.2.63
1013
1014- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1015
1016## 0.2.61
1017
1018- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1019- Enhanced image detection for more reliable clipboard paste functionality
1020- Fixed an issue where ESC key could crash the conversation history selector
1021
1022## 0.2.59
1023
1024- Copy+paste images directly into your prompt
1025- Improved progress indicators for bash and fetch tools
1026- Bugfixes for non-interactive mode (-p)
1027
1028## 0.2.54
1029
1030- Quickly add to Memory by starting your message with '#'
1031- Press ctrl+r to see full output for long tool results
1032- Added support for MCP SSE transport
1033
1034## 0.2.53
1035
1036- New web fetch tool lets Claude view URLs that you paste in
1037- Fixed a bug with JPEG detection
1038
1039## 0.2.50
1040
1041- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1042
1043## 0.2.49
1044
1045- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1046
1047## 0.2.47
1048
1049- Press Tab to auto-complete file and folder names
1050- Press Shift + Tab to toggle auto-accept for file edits
1051- Automatic conversation compaction for infinite conversation length (toggle with /config)
1052
1053## 0.2.44
1054
1055- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1056
1057## 0.2.41
1058
1059- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1060- MCP server startup no longer blocks the app from starting up
1061
1062## 0.2.37
1063
1064- New /release-notes command lets you view release notes at any time
1065- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1066
1067## 0.2.36
1068
1069- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1070- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1071
1072## 0.2.34
1073
1074- Vim bindings for text input - enable with /vim or /config
1075
1076## 0.2.32
1077
1078- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1079- Fix for some PersistentShell issues
1080
1081## 0.2.31
1082
1083- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1084- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1085
1086## 0.2.30
1087
1088- Added ANSI color theme for better terminal compatibility
1089- Fixed issue where slash command arguments weren't being sent properly
1090- (Mac-only) API keys are now stored in macOS Keychain
1091
1092## 0.2.26
1093
1094- New /approved-tools command for managing tool permissions
1095- Word-level diff display for improved code readability
1096- Fuzzy matching for slash commands
1097
1098## 0.2.21
1099
1100- Fuzzy matching for /commands