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 eb87245 · on Dec 13, 2025 History
1,112 lines · 44.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.69
4
5- Minor bugfixes
6
7## 2.0.68
8
9- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
10- Fixed a bug where disallowed MCP tools were visible to the model
11- Fixed an issue where steering messages could be lost while a subagent is working
12- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
13- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
14- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
15
16## 2.0.67
17
18- Thinking mode is now enabled by default for Opus 4.5
19- Thinking mode configuration has moved to /config
20- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
21- Show reason why autoupdater is disabled in `/doctor`
22- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
23- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
24- Fixed scroll position resetting after deleting a permission rule in `/permissions`
25- 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
26- Fixed `claude install --force` not bypassing stale lock files
27- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
28- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
29
30## 2.0.65
31
32- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
33- Added context window information to status line input
34- Added `fileSuggestion` setting for custom `@` file search commands
35- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
36- Fixed prompt not being saved to history when aborting a query with Escape
37- Fixed Read tool image handling to identify format from bytes instead of file extension
38
39## 2.0.64
40
41- Made auto-compacting instant
42- Agents and bash commands can run asynchronously and send messages to wake up the main agent
43- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
44- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
45- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
46- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
47- Fixed auto-loading .env when using native installer
48- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
49- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
50- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
51- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
52- Bedrock: Improve efficiency of token counting
53- Bedrock: Add support for `aws login` AWS Management Console credentials
54- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
55
56## 2.0.62
57
58- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
59- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
60- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
61- Fixed slash command selection not working when multiple commands share the same name
62- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
63- Fixed running versions getting removed because lock file incorrectly going stale
64- Fixed IDE diff tab not closing when rejecting file changes
65
66## 2.0.61
67
68- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
69
70## 2.0.60
71
72- Added background agent support. Agents run in the background while you work
73- Added --disable-slash-commands CLI flag to disable all slash commands
74- Added model name to "Co-Authored-By" commit messages
75- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
76- Updated Fetch to skip summarization for pre-approved websites
77- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
78
79## 2.0.59
80
81- Added --agent CLI flag to override the agent setting for the current session
82- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
83- VS Code: Fixed .claude.json config file being read from incorrect location
84
85## 2.0.58
86
87- Pro users now have access to Opus 4.5 as part of their subscription!
88- Fixed timer duration showing "11m 60s" instead of "12m 0s"
89- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
90
91## 2.0.57
92
93- Added feedback input when rejecting plans, allowing users to tell Claude what to change
94- VSCode: Added streaming message support for real-time response display
95
96## 2.0.56
97
98- Added setting to enable/disable terminal progress bar (OSC 9;4)
99- 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.
100
101## 2.0.55
102
103- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
104- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
105- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
106- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
107
108## 2.0.54
109
110- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
111- Fix issue with excessive iTerm notifications
112
113## 2.0.52
114
115- Fixed duplicate message display when starting Claude with a command line argument
116- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
117- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
118- Permit some uses of `$!` in bash commands
119
120## 2.0.51
121
122- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
123- Introducing Claude Code for Desktop: https://claude.com/download
124- 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
125- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
126- Plan Mode now builds more precise plans and executes more thoroughly
127- Usage limit notifications now easier to understand
128- Switched `/usage` back to "% used"
129- Fixed handling of thinking errors
130- Fixed performance regression
131
132## 2.0.50
133
134- Fixed bug preventing calling MCP tools that have nested references in their input schemas
135- Silenced a noisy but harmless error during upgrades
136- Improved ultrathink text display
137- Improved clarity of 5-hour session limit warning message
138
139## 2.0.49
140
141- Added readline-style ctrl-y for pasting deleted text
142- Improved clarity of usage limit warning message
143- Fixed handling of subagent permissions
144
145## 2.0.47
146
147- Improved error messages and validation for `claude --teleport`
148- Improved error handling in `/usage`
149- Fixed race condition with history entry not getting logged at exit
150- Fixed Vertex AI configuration not being applied from `settings.json`
151
152## 2.0.46
153
154- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
155
156## 2.0.45
157
158- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
159- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
160- Send background tasks to Claude Code on the web by starting a message with `&`
161
162## 2.0.43
163
164- Added `permissionMode` field for custom agents
165- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
166- Added skills frontmatter field to declare skills to auto-load for subagents
167- Added the `SubagentStart` hook event
168- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
169- Fixed duplicate rendering of some messages in the UI
170- Fixed some visual flickers
171- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
172
173## 2.0.42
174
175- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
176
177## 2.0.41
178
179- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
180- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
181- Fixed incorrect labeling of user settings vs project settings in command descriptions
182- Fixed crash when plugin command hooks timeout during execution
183- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
184- Fixed broken security documentation links in trust dialogs and onboarding
185- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
186- ctrl-r history search landing on a slash command no longer cancels the search
187- SDK: Support custom timeouts for hooks
188- Allow more safe git commands to run without approval
189- Plugins: Added support for sharing and installing output styles
190- Teleporting a session from web will automatically set the upstream branch
191
192## 2.0.37
193
194- Fixed how idleness is computed for notifications
195- Hooks: Added matcher values for Notification hook events
196- Output Styles: Added `keep-coding-instructions` option to frontmatter
197
198## 2.0.36
199
200- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
201- Fixed queued messages being incorrectly executed as bash commands
202- Fixed input being lost when typing while a queued message is processed
203
204## 2.0.35
205
206- Improve fuzzy search results when searching commands
207- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
208- 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
209- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
210- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
211
212## 2.0.34
213
214- VSCode Extension: Added setting to configure the initial permission mode for new conversations
215- Improved file path suggestion performance with native Rust-based fuzzy finder
216- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
217- Fixed memory crash when reading or writing large files (especially base64-encoded images)
218
219## 2.0.33
220
221- Native binary installs now launch quicker.
222- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
223- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
224
225## 2.0.32
226
227- Un-deprecate output styles based on community feedback
228- Added `companyAnnouncements` setting for displaying announcements on startup
229- Fixed hook progress messages not updating correctly during PostToolUse hook execution
230
231## 2.0.31
232
233- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
234- Vertex: add support for Web Search on supported models
235- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
236- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
237- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
238- Fixed plugin uninstall not removing plugins
239
240## 2.0.30
241
242- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
243- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
244- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
245- Added prompt-based stop hooks
246- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
247- Enabled SSE MCP servers on native build
248- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
249- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
250- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
251- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
252- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
253- Fixed bug that saved session permissions to local settings
254- Fixed MCP tools not being available to sub-agents
255- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
256- Fixed delay when navigating through typeahead suggestions with arrow keys
257- VSCode: Restored selection indicator in input footer showing current file or code selection status
258
259## 2.0.28
260
261- Plan mode: introduced new Plan subagent
262- Subagents: claude can now choose to resume subagents
263- Subagents: claude can dynamically choose the model used by its subagents
264- SDK: added --max-budget-usd flag
265- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
266- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
267- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
268- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
269- Various other bug fixes
270
271## 2.0.27
272
273- New UI for permission prompts
274- Added current branch filtering and search to session resume screen for easier navigation
275- Fixed directory @-mention causing "No assistant message found" error
276- VSCode Extension: Add config setting to include .gitignored files in file searches
277- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
278
279## 2.0.25
280
281- 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
282
283## 2.0.24
284
285- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
286- Claude Code Web: Support for Web -> CLI teleport
287- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
288- Bedrock: Display awsAuthRefresh output when auth is required
289
290## 2.0.22
291
292- Fixed content layout shift when scrolling through slash commands
293- IDE: Add toggle to enable/disable thinking.
294- Fix bug causing duplicate permission prompts with parallel tool calls
295- Add support for enterprise managed MCP allowlist and denylist
296
297## 2.0.21
298
299- Support MCP `structuredContent` field in tool responses
300- Added an interactive question tool
301- Claude will now ask you questions more often in plan mode
302- Added Haiku 4.5 as a model option for Pro users
303- Fixed an issue where queued commands don't have access to previous messages' output
304
305## 2.0.20
306
307- Added support for Claude Skills
308
309## 2.0.19
310
311- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
312- Fixed a bug where Haiku was unnecessarily called in print mode
313
314## 2.0.17
315
316- Added Haiku 4.5 to model selector!
317- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
318- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
319- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
320- OTEL: support HTTP_PROXY and HTTPS_PROXY
321- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
322
323## 2.0.15
324
325- Fixed bug with resuming where previously created files needed to be read again before writing
326- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
327
328## 2.0.14
329
330- Fix @-mentioning MCP servers to toggle them on/off
331- Improve permission checks for bash with inline env vars
332- Fix ultrathink + thinking toggle
333- Reduce unnecessary logins
334- Document --system-prompt
335- Several improvements to rendering
336- Plugins UI polish
337
338## 2.0.13
339
340- Fixed `/plugin` not working on native build
341
342## 2.0.12
343
344- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
345- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
346- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
347- `/plugin validate` command for validating plugin structure and configuration
348- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
349- Plugin documentation available at https://code.claude.com/docs/en/plugins
350- Comprehensive error messages and diagnostics via `/doctor` command
351- Avoid flickering in `/model` selector
352- Improvements to `/help`
353- Avoid mentioning hooks in `/resume` summaries
354- Changes to the "verbose" setting in `/config` now persist across sessions
355
356## 2.0.11
357
358- Reduced system prompt size by 1.4k tokens
359- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
360- Fixed Opus fallback rate limit errors appearing incorrectly
361- Fixed /add-dir command selecting wrong default tab
362
363## 2.0.10
364
365- Rewrote terminal renderer for buttery smooth UI
366- Enable/disable MCP servers by @mentioning, or in /mcp
367- Added tab completion for shell commands in bash mode
368- PreToolUse hooks can now modify tool inputs
369- Press Ctrl-G to edit your prompt in your system's configured text editor
370- Fixes for bash permission checks with environment variables in the command
371
372## 2.0.9
373
374- Fix regression where bash backgrounding stopped working
375
376## 2.0.8
377
378- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
379- IDE: Add drag-and-drop support for files and folders in chat
380- /context: Fix counting for thinking blocks
381- Improve message rendering for users with light themes on dark terminals
382- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
383
384## 2.0.5
385
386- IDE: Fix IME unintended message submission with Enter and Tab
387- IDE: Add "Open in Terminal" link in login screen
388- Fix unhandled OAuth expiration 401 API errors
389- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
390
391## 2.0.1
392
393- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
394- Various bug fixes and presentation improvements
395
396## 2.0.0
397
398- New native VS Code extension
399- Fresh coat of paint throughout the whole app
400- /rewind a conversation to undo code changes
401- /usage command to see plan limits
402- Tab to toggle thinking (sticky across sessions)
403- Ctrl-R to search history
404- Unshipped claude config command
405- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
406- SDK: The Claude Code SDK is now the Claude Agent SDK
407- Add subagents dynamically with `--agents` flag
408
409## 1.0.126
410
411- Enable /context command for Bedrock and Vertex
412- Add mTLS support for HTTP-based OpenTelemetry exporters
413
414## 1.0.124
415
416- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
417- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
418- No longer inform Claude of the list of allowed tools when permission is denied
419- Fixed security vulnerability in Bash tool permission checks
420- Improved VSCode extension performance for large files
421
422## 1.0.123
423
424- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
425- Fixed thinking mode triggering on negation phrases like "don't think"
426- Fixed rendering performance degradation during token streaming
427- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
428- Enhanced BashTool environment snapshot logging
429- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
430- Migrated --debug logging to a file, to enable easy tailing & filtering
431
432## 1.0.120
433
434- Fix input lag during typing, especially noticeable with large prompts
435- Improved VSCode extension command registry and sessions dialog user experience
436- Enhanced sessions dialog responsiveness and visual feedback
437- Fixed IDE compatibility issue by removing worktree support check
438- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
439
440## 1.0.119
441
442- Fix Windows issue where process visually freezes on entering interactive mode
443- Support dynamic headers for MCP servers via headersHelper configuration
444- Fix thinking mode not working in headless sessions
445- Fix slash commands now properly update allowed tools instead of replacing them
446
447## 1.0.117
448
449- Add Ctrl-R history search to recall previous commands like bash/zsh
450- Fix input lag while typing, especially on Windows
451- Add sed command to auto-allowed commands in acceptEdits mode
452- Fix Windows PATH comparison to be case-insensitive for drive letters
453- Add permissions management hint to /add-dir output
454
455## 1.0.115
456
457- Improve thinking mode display with enhanced visual effects
458- Type /t to temporarily disable thinking mode in your prompt
459- Improve path validation for glob and grep tools
460- Show condensed output for post-tool hooks to reduce visual clutter
461- Fix visual feedback when loading state completes
462- Improve UI consistency for permission request dialogs
463
464## 1.0.113
465
466- Deprecated piped input in interactive mode
467- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
468
469## 1.0.112
470
471- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
472- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
473- Hooks: Added systemMessage support for SessionEnd hooks
474- Added `spinnerTipsEnabled` setting to disable spinner tips
475- IDE: Various improvements and bug fixes
476
477## 1.0.111
478
479- /model now validates provided model names
480- Fixed Bash tool crashes caused by malformed shell syntax parsing
481
482## 1.0.110
483
484- /terminal-setup command now supports WezTerm
485- MCP: OAuth tokens now proactively refresh before expiration
486- Fixed reliability issues with background Bash processes
487
488## 1.0.109
489
490- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
491
492## 1.0.106
493
494- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
495
496## 1.0.97
497
498- Settings: /doctor now validates permission rule syntax and suggests corrections
499
500## 1.0.94
501
502- Vertex: add support for global endpoints for supported models
503- /memory command now allows direct editing of all imported memory files
504- SDK: Add custom tools as callbacks
505- Added /todos command to list current todo items
506
507## 1.0.93
508
509- Windows: Add alt + v shortcut for pasting images from clipboard
510- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
511
512## 1.0.90
513
514- Settings file changes take effect immediately - no restart required
515
516## 1.0.88
517
518- Fixed issue causing "OAuth authentication is currently not supported"
519- Status line input now includes `exceeds_200k_tokens`
520- Fixed incorrect usage tracking in /cost.
521- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
522- Bedrock: Updated default Sonnet model to Sonnet 4
523
524## 1.0.86
525
526- Added /context to help users self-serve debug context issues
527- SDK: Added UUID support for all SDK messages
528- SDK: Added `--replay-user-messages` to replay user messages back to stdout
529
530## 1.0.85
531
532- Status line input now includes session cost info
533- Hooks: Introduced SessionEnd hook
534
535## 1.0.84
536
537- Fix tool_use/tool_result id mismatch error when network is unstable
538- Fix Claude sometimes ignoring real-time steering when wrapping up a task
539- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
540- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
541
542## 1.0.83
543
544- @-mention: Support files with spaces in path
545- New shimmering spinner
546
547## 1.0.82
548
549- SDK: Add request cancellation support
550- SDK: New additionalDirectories option to search custom paths, improved slash command processing
551- Settings: Validation prevents invalid fields in .claude/settings.json files
552- MCP: Improve tool name consistency
553- Bash: Fix crash when Claude tries to automatically read large files
554
555## 1.0.81
556
557- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
558- Agents: Fix custom agent loading when agent files are unparsable
559
560## 1.0.80
561
562- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
563
564## 1.0.77
565
566- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
567- SDK: Add session support and permission denial tracking
568- Fix token limit errors in conversation summarization
569- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
570
571## 1.0.73
572
573- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
574- MCP: Press Esc to cancel OAuth authentication flows
575- Bash: Improved command validation and reduced false security warnings
576- UI: Enhanced spinner animations and status line visual hierarchy
577- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
578
579## 1.0.72
580
581- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
582
583## 1.0.71
584
585- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
586- Customizable status line: add your terminal prompt to Claude Code with /statusline
587
588## 1.0.70
589
590- Performance: Optimized message rendering for better performance with large contexts
591- Windows: Fixed native file search, ripgrep, and subagent functionality
592- Added support for @-mentions in slash command arguments
593
594## 1.0.69
595
596- Upgraded Opus to version 4.1
597
598## 1.0.68
599
600- Fix incorrect model names being used for certain commands like `/pr-comments`
601- 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.
602- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
603- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
604- SDK: Added canUseTool callback support for tool confirmation
605- Added `disableAllHooks` setting
606- Improved file suggestions performance in large repos
607
608## 1.0.65
609
610- IDE: Fixed connection stability issues and error handling for diagnostics
611- Windows: Fixed shell environment setup for users without .bashrc files
612
613## 1.0.64
614
615- Agents: Added model customization support - you can now specify which model an agent should use
616- Agents: Fixed unintended access to the recursive agent tool
617- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
618- SDK: Fixed user input tracking across multi-turn conversations
619- Added hidden files to file search and @-mention suggestions
620
621## 1.0.63
622
623- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
624
625## 1.0.62
626
627- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
628- Hooks: Added SessionStart hook for new session initialization
629- /add-dir command now supports typeahead for directory paths
630- Improved network connectivity check reliability
631
632## 1.0.61
633
634- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
635- Settings: Added `--settings` flag to load settings from a JSON file
636- Settings: Fixed resolution of settings files paths that are symlinks
637- OTEL: Fixed reporting of wrong organization after authentication changes
638- Slash commands: Fixed permissions checking for allowed-tools with Bash
639- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
640- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
641- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
642
643## 1.0.60
644
645- You can now create custom subagents for specialized tasks! Run /agents to get started
646
647## 1.0.59
648
649- SDK: Added tool confirmation support with canUseTool callback
650- SDK: Allow specifying env for spawned process
651- Hooks: Exposed PermissionDecision to hooks (including "ask")
652- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
653- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
654
655## 1.0.58
656
657- Added support for reading PDFs
658- MCP: Improved server health status display in 'claude mcp list'
659- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
660
661## 1.0.57
662
663- Added support for specifying a model in slash commands
664- Improved permission messages to help Claude understand allowed tools
665- Fix: Remove trailing newlines from bash output in terminal wrapping
666
667## 1.0.56
668
669- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
670- Fixes for WSL IDE detection
671- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
672
673## 1.0.55
674
675- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
676- Windows: fixed Ctrl+Z crash
677- SDK: Added ability to capture error logging
678- Add --system-prompt-file option to override system prompt in print mode
679
680## 1.0.54
681
682- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
683- Custom slash commands: Added argument-hint to frontmatter
684- Windows: OAuth uses port 45454 and properly constructs browser URL
685- Windows: mode switching now uses alt + m, and plan mode renders properly
686- Shell: Switch to in-memory shell snapshot to fix file-related errors
687
688## 1.0.53
689
690- Updated @-mention file truncation from 100 lines to 2000 lines
691- 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).
692
693## 1.0.52
694
695- Added support for MCP server instructions
696
697## 1.0.51
698
699- Added support for native Windows (requires Git for Windows)
700- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
701- Settings: /doctor can now help you identify and fix invalid setting files
702- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
703- Increased auto-compact warning threshold from 60% to 80%
704- Fixed an issue with handling user directories with spaces for shell snapshots
705- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
706- Custom slash commands: Fixed user-level commands in subdirectories
707- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
708
709## 1.0.48
710
711- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
712- Added progress messages to Bash tool based on the last 5 lines of command output
713- Added expanding variables support for MCP server configuration
714- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
715- Improved IDE extension path handling when Claude Code runs in WSL
716- Hooks: Added a PreCompact hook
717- Vim mode: Added c, f/F, t/T
718
719## 1.0.45
720
721- Redesigned Search (Grep) tool with new tool input parameters and features
722- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
723- Fixed config file corruption issue by enforcing atomic writes
724- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
725- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
726- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
727
728## 1.0.44
729
730- New /export command lets you quickly export a conversation for sharing
731- MCP: resource_link tool results are now supported
732- MCP: tool annotations and tool titles now display in /mcp view
733- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
734
735## 1.0.43
736
737- Fixed a bug where the theme selector was saving excessively
738- Hooks: Added EPIPE system error handling
739
740## 1.0.42
741
742- Added tilde (`~`) expansion support to `/add-dir` command
743
744## 1.0.41
745
746- Hooks: Split Stop hook triggering into Stop and SubagentStop
747- Hooks: Enabled optional timeout configuration for each command
748- Hooks: Added "hook_event_name" to hook input
749- Fixed a bug where MCP tools would display twice in tool list
750- New tool parameters JSON for Bash tool in `tool_decision` event
751
752## 1.0.40
753
754- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
755
756## 1.0.39
757
758- New Active Time metric in OpenTelemetry logging
759
760## 1.0.38
761
762- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
763
764## 1.0.37
765
766- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
767
768## 1.0.36
769
770- Web search now takes today's date into context
771- Fixed a bug where stdio MCP servers were not terminating properly on exit
772
773## 1.0.35
774
775- Added support for MCP OAuth Authorization Server discovery
776
777## 1.0.34
778
779- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
780
781## 1.0.33
782
783- Improved logging functionality with session ID support
784- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
785- Improvements to plan mode
786
787## 1.0.32
788
789- Updated loopback config for litellm
790- Added forceLoginMethod setting to bypass login selection screen
791
792## 1.0.31
793
794- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
795
796## 1.0.30
797
798- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
799- Improved file path autocomplete with filename matching
800- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
801- Enhanced jq regex support for complex filters with pipes and select
802
803## 1.0.29
804
805- Improved CJK character support in cursor navigation and rendering
806
807## 1.0.28
808
809- Slash commands: Fix selector display during history navigation
810- Resizes images before upload to prevent API size limit errors
811- Added XDG_CONFIG_HOME support to configuration directory
812- Performance optimizations for memory usage
813- New attributes (terminal.type, language) in OpenTelemetry logging
814
815## 1.0.27
816
817- Streamable HTTP MCP servers are now supported
818- Remote MCP servers (SSE and HTTP) now support OAuth
819- MCP resources can now be @-mentioned
820- /resume slash command to switch conversations within Claude Code
821
822## 1.0.25
823
824- Slash commands: moved "project" and "user" prefixes to descriptions
825- Slash commands: improved reliability for command discovery
826- Improved support for Ghostty
827- Improved web search reliability
828
829## 1.0.24
830
831- Improved /mcp output
832- Fixed a bug where settings arrays got overwritten instead of merged
833
834## 1.0.23
835
836- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
837- Released Python SDK: pip install claude-code-sdk to get started
838
839## 1.0.22
840
841- SDK: Renamed `total_cost` to `total_cost_usd`
842
843## 1.0.21
844
845- Improved editing of files with tab-based indentation
846- Fix for tool_use without matching tool_result errors
847- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
848
849## 1.0.18
850
851- Added --add-dir CLI argument for specifying additional working directories
852- Added streaming input support without require -p flag
853- Improved startup performance and session storage performance
854- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
855- Added detailed MCP server tools display (/mcp)
856- MCP authentication and permission improvements
857- Added auto-reconnection for MCP SSE connections on disconnect
858- Fixed issue where pasted content was lost when dialogs appeared
859
860## 1.0.17
861
862- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
863- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
864- MCP server list UI improvements
865- Update Claude Code process title to display "claude" instead of "node"
866
867## 1.0.11
868
869- Claude Code can now also be used with a Claude Pro subscription
870- Added /upgrade for smoother switching to Claude Max plans
871- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
872- Improved shell configuration error handling
873- Improved todo list handling during compaction
874
875## 1.0.10
876
877- Added markdown table support
878- Improved streaming performance
879
880## 1.0.8
881
882- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
883- Increased default otel interval from 1s -> 5s
884- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
885- Fixed a regression where search tools unnecessarily asked for permissions
886- Added support for triggering thinking non-English languages
887- Improved compacting UI
888
889## 1.0.7
890
891- Renamed /allowed-tools -> /permissions
892- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
893- Deprecated claude config commands in favor of editing settings.json
894- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
895- Improved error handling for /install-github-app
896- Bugfixes, UI polish, and tool reliability improvements
897
898## 1.0.6
899
900- Improved edit reliability for tab-indented files
901- Respect CLAUDE_CONFIG_DIR everywhere
902- Reduced unnecessary tool permission prompts
903- Added support for symlinks in @file typeahead
904- Bugfixes, UI polish, and tool reliability improvements
905
906## 1.0.4
907
908- Fixed a bug where MCP tool errors weren't being parsed correctly
909
910## 1.0.1
911
912- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
913- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
914- Updated documentation links and OAuth process descriptions
915
916## 1.0.0
917
918- Claude Code is now generally available
919- Introducing Sonnet 4 and Opus 4 models
920
921## 0.2.125
922
923- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
924- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
925
926## 0.2.117
927
928- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
929- Introduced settings.cleanupPeriodDays
930- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
931- Introduced --debug mode
932
933## 0.2.108
934
935- You can now send messages to Claude while it works to steer Claude in real-time
936- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
937- Fixed a bug where thinking was not working in -p mode
938- Fixed a regression in /cost reporting
939- Deprecated MCP wizard interface in favor of other MCP commands
940- Lots of other bugfixes and improvements
941
942## 0.2.107
943
944- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
945
946## 0.2.106
947
948- MCP SSE server configs can now specify custom headers
949- Fixed a bug where MCP permission prompt didn't always show correctly
950
951## 0.2.105
952
953- Claude can now search the web
954- Moved system & account status to /status
955- Added word movement keybindings for Vim
956- Improved latency for startup, todo tool, and file edits
957
958## 0.2.102
959
960- Improved thinking triggering reliability
961- Improved @mention reliability for images and folders
962- You can now paste multiple large chunks into one prompt
963
964## 0.2.100
965
966- Fixed a crash caused by a stack overflow error
967- Made db storage optional; missing db support disables --continue and --resume
968
969## 0.2.98
970
971- Fixed an issue where auto-compact was running twice
972
973## 0.2.96
974
975- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
976
977## 0.2.93
978
979- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
980- Claude now has access to a Todo list that helps it stay on track and be more organized
981
982## 0.2.82
983
984- Added support for --disallowedTools
985- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
986
987## 0.2.75
988
989- Hit Enter to queue up additional messages while Claude is working
990- Drag in or copy/paste image files directly into the prompt
991- @-mention files to directly add them to context
992- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
993- Improved performance for filename auto-complete
994
995## 0.2.74
996
997- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
998- Task tool can now perform writes and run bash commands
999
1000## 0.2.72
1001
1002- Updated spinner to indicate tokens loaded and tool usage
1003
1004## 0.2.70
1005
1006- Network commands like curl are now available for Claude to use
1007- Claude can now run multiple web queries in parallel
1008- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1009
1010## 0.2.69
1011
1012- Fixed UI glitches with improved Select component behavior
1013- Enhanced terminal output display with better text truncation logic
1014
1015## 0.2.67
1016
1017- Shared project permission rules can be saved in .claude/settings.json
1018
1019## 0.2.66
1020
1021- Print mode (-p) now supports streaming output via --output-format=stream-json
1022- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1023
1024## 0.2.63
1025
1026- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1027
1028## 0.2.61
1029
1030- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1031- Enhanced image detection for more reliable clipboard paste functionality
1032- Fixed an issue where ESC key could crash the conversation history selector
1033
1034## 0.2.59
1035
1036- Copy+paste images directly into your prompt
1037- Improved progress indicators for bash and fetch tools
1038- Bugfixes for non-interactive mode (-p)
1039
1040## 0.2.54
1041
1042- Quickly add to Memory by starting your message with '#'
1043- Press ctrl+r to see full output for long tool results
1044- Added support for MCP SSE transport
1045
1046## 0.2.53
1047
1048- New web fetch tool lets Claude view URLs that you paste in
1049- Fixed a bug with JPEG detection
1050
1051## 0.2.50
1052
1053- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1054
1055## 0.2.49
1056
1057- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1058
1059## 0.2.47
1060
1061- Press Tab to auto-complete file and folder names
1062- Press Shift + Tab to toggle auto-accept for file edits
1063- Automatic conversation compaction for infinite conversation length (toggle with /config)
1064
1065## 0.2.44
1066
1067- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1068
1069## 0.2.41
1070
1071- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1072- MCP server startup no longer blocks the app from starting up
1073
1074## 0.2.37
1075
1076- New /release-notes command lets you view release notes at any time
1077- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1078
1079## 0.2.36
1080
1081- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1082- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1083
1084## 0.2.34
1085
1086- Vim bindings for text input - enable with /vim or /config
1087
1088## 0.2.32
1089
1090- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1091- Fix for some PersistentShell issues
1092
1093## 0.2.31
1094
1095- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1096- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1097
1098## 0.2.30
1099
1100- Added ANSI color theme for better terminal compatibility
1101- Fixed issue where slash command arguments weren't being sent properly
1102- (Mac-only) API keys are now stored in macOS Keychain
1103
1104## 0.2.26
1105
1106- New /approved-tools command for managing tool permissions
1107- Word-level diff display for improved code readability
1108- Fuzzy matching for slash commands
1109
1110## 0.2.21
1111
1112- Fuzzy matching for /commands