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 de49a07 · on Dec 7, 2025 History
1,049 lines · 40.7 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.61
4
5- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
6
7## 2.0.60
8
9- Added background agent support. Agents run in the background while you work
10- Added --disable-slash-commands CLI flag to disable all slash commands
11- Added model name to "Co-Authored-By" commit messages
12- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
13- Updated Fetch to skip summarization for pre-approved websites
14- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
15
16## 2.0.59
17
18- Added --agent CLI flag to override the agent setting for the current session
19- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
20- VS Code: Fixed .claude.json config file being read from incorrect location
21
22## 2.0.58
23
24- Pro users now have access to Opus 4.5 as part of their subscription!
25- Fixed timer duration showing "11m 60s" instead of "12m 0s"
26- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
27
28## 2.0.57
29
30- Added feedback input when rejecting plans, allowing users to tell Claude what to change
31- VSCode: Added streaming message support for real-time response display
32
33## 2.0.56
34
35- Added setting to enable/disable terminal progress bar (OSC 9;4)
36- 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.
37
38## 2.0.55
39
40- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
41- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
42- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
43- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
44
45## 2.0.54
46
47- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
48- Fix issue with excessive iTerm notifications
49
50## 2.0.52
51
52- Fixed duplicate message display when starting Claude with a command line argument
53- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
54- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
55- Permit some uses of `$!` in bash commands
56
57## 2.0.51
58
59- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
60- Introducing Claude Code for Desktop: https://claude.com/download
61- 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
62- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
63- Plan Mode now builds more precise plans and executes more thoroughly
64- Usage limit notifications now easier to understand
65- Switched `/usage` back to "% used"
66- Fixed handling of thinking errors
67- Fixed performance regression
68
69## 2.0.50
70
71- Fixed bug preventing calling MCP tools that have nested references in their input schemas
72- Silenced a noisy but harmless error during upgrades
73- Improved ultrathink text display
74- Improved clarity of 5-hour session limit warning message
75
76## 2.0.49
77
78- Added readline-style ctrl-y for pasting deleted text
79- Improved clarity of usage limit warning message
80- Fixed handling of subagent permissions
81
82## 2.0.47
83
84- Improved error messages and validation for `claude --teleport`
85- Improved error handling in `/usage`
86- Fixed race condition with history entry not getting logged at exit
87- Fixed Vertex AI configuration not being applied from `settings.json`
88
89## 2.0.46
90
91- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
92
93## 2.0.45
94
95- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
96- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
97- Send background tasks to Claude Code on the web by starting a message with `&`
98
99## 2.0.43
100
101- Added `permissionMode` field for custom agents
102- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
103- Added skills frontmatter field to declare skills to auto-load for subagents
104- Added the `SubagentStart` hook event
105- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
106- Fixed duplicate rendering of some messages in the UI
107- Fixed some visual flickers
108- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
109
110## 2.0.42
111
112- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
113
114## 2.0.41
115
116- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
117- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
118- Fixed incorrect labeling of user settings vs project settings in command descriptions
119- Fixed crash when plugin command hooks timeout during execution
120- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
121- Fixed broken security documentation links in trust dialogs and onboarding
122- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
123- ctrl-r history search landing on a slash command no longer cancels the search
124- SDK: Support custom timeouts for hooks
125- Allow more safe git commands to run without approval
126- Plugins: Added support for sharing and installing output styles
127- Teleporting a session from web will automatically set the upstream branch
128
129## 2.0.37
130
131- Fixed how idleness is computed for notifications
132- Hooks: Added matcher values for Notification hook events
133- Output Styles: Added `keep-coding-instructions` option to frontmatter
134
135## 2.0.36
136
137- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
138- Fixed queued messages being incorrectly executed as bash commands
139- Fixed input being lost when typing while a queued message is processed
140
141## 2.0.35
142
143- Improve fuzzy search results when searching commands
144- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
145- 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
146- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
147- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
148
149## 2.0.34
150
151- VSCode Extension: Added setting to configure the initial permission mode for new conversations
152- Improved file path suggestion performance with native Rust-based fuzzy finder
153- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
154- Fixed memory crash when reading or writing large files (especially base64-encoded images)
155
156## 2.0.33
157
158- Native binary installs now launch quicker.
159- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
160- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
161
162## 2.0.32
163
164- Un-deprecate output styles based on community feedback
165- Added `companyAnnouncements` setting for displaying announcements on startup
166- Fixed hook progress messages not updating correctly during PostToolUse hook execution
167
168## 2.0.31
169
170- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
171- Vertex: add support for Web Search on supported models
172- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
173- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
174- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
175- Fixed plugin uninstall not removing plugins
176
177## 2.0.30
178
179- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
180- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
181- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
182- Added prompt-based stop hooks
183- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
184- Enabled SSE MCP servers on native build
185- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
186- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
187- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
188- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
189- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
190- Fixed bug that saved session permissions to local settings
191- Fixed MCP tools not being available to sub-agents
192- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
193- Fixed delay when navigating through typeahead suggestions with arrow keys
194- VSCode: Restored selection indicator in input footer showing current file or code selection status
195
196## 2.0.28
197
198- Plan mode: introduced new Plan subagent
199- Subagents: claude can now choose to resume subagents
200- Subagents: claude can dynamically choose the model used by its subagents
201- SDK: added --max-budget-usd flag
202- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
203- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
204- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
205- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
206- Various other bug fixes
207
208## 2.0.27
209
210- New UI for permission prompts
211- Added current branch filtering and search to session resume screen for easier navigation
212- Fixed directory @-mention causing "No assistant message found" error
213- VSCode Extension: Add config setting to include .gitignored files in file searches
214- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
215
216## 2.0.25
217
218- 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
219
220## 2.0.24
221
222- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
223- Claude Code Web: Support for Web -> CLI teleport
224- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
225- Bedrock: Display awsAuthRefresh output when auth is required
226
227## 2.0.22
228
229- Fixed content layout shift when scrolling through slash commands
230- IDE: Add toggle to enable/disable thinking.
231- Fix bug causing duplicate permission prompts with parallel tool calls
232- Add support for enterprise managed MCP allowlist and denylist
233
234## 2.0.21
235
236- Support MCP `structuredContent` field in tool responses
237- Added an interactive question tool
238- Claude will now ask you questions more often in plan mode
239- Added Haiku 4.5 as a model option for Pro users
240- Fixed an issue where queued commands don't have access to previous messages' output
241
242## 2.0.20
243
244- Added support for Claude Skills
245
246## 2.0.19
247
248- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
249- Fixed a bug where Haiku was unnecessarily called in print mode
250
251## 2.0.17
252
253- Added Haiku 4.5 to model selector!
254- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
255- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
256- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
257- OTEL: support HTTP_PROXY and HTTPS_PROXY
258- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
259
260## 2.0.15
261
262- Fixed bug with resuming where previously created files needed to be read again before writing
263- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
264
265## 2.0.14
266
267- Fix @-mentioning MCP servers to toggle them on/off
268- Improve permission checks for bash with inline env vars
269- Fix ultrathink + thinking toggle
270- Reduce unnecessary logins
271- Document --system-prompt
272- Several improvements to rendering
273- Plugins UI polish
274
275## 2.0.13
276
277- Fixed `/plugin` not working on native build
278
279## 2.0.12
280
281- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
282- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
283- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
284- `/plugin validate` command for validating plugin structure and configuration
285- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
286- Plugin documentation available at https://code.claude.com/docs/en/plugins
287- Comprehensive error messages and diagnostics via `/doctor` command
288- Avoid flickering in `/model` selector
289- Improvements to `/help`
290- Avoid mentioning hooks in `/resume` summaries
291- Changes to the "verbose" setting in `/config` now persist across sessions
292
293## 2.0.11
294
295- Reduced system prompt size by 1.4k tokens
296- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
297- Fixed Opus fallback rate limit errors appearing incorrectly
298- Fixed /add-dir command selecting wrong default tab
299
300## 2.0.10
301
302- Rewrote terminal renderer for buttery smooth UI
303- Enable/disable MCP servers by @mentioning, or in /mcp
304- Added tab completion for shell commands in bash mode
305- PreToolUse hooks can now modify tool inputs
306- Press Ctrl-G to edit your prompt in your system's configured text editor
307- Fixes for bash permission checks with environment variables in the command
308
309## 2.0.9
310
311- Fix regression where bash backgrounding stopped working
312
313## 2.0.8
314
315- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
316- IDE: Add drag-and-drop support for files and folders in chat
317- /context: Fix counting for thinking blocks
318- Improve message rendering for users with light themes on dark terminals
319- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
320
321## 2.0.5
322
323- IDE: Fix IME unintended message submission with Enter and Tab
324- IDE: Add "Open in Terminal" link in login screen
325- Fix unhandled OAuth expiration 401 API errors
326- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
327
328## 2.0.1
329
330- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
331- Various bug fixes and presentation improvements
332
333## 2.0.0
334
335- New native VS Code extension
336- Fresh coat of paint throughout the whole app
337- /rewind a conversation to undo code changes
338- /usage command to see plan limits
339- Tab to toggle thinking (sticky across sessions)
340- Ctrl-R to search history
341- Unshipped claude config command
342- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
343- SDK: The Claude Code SDK is now the Claude Agent SDK
344- Add subagents dynamically with `--agents` flag
345
346## 1.0.126
347
348- Enable /context command for Bedrock and Vertex
349- Add mTLS support for HTTP-based OpenTelemetry exporters
350
351## 1.0.124
352
353- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
354- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
355- No longer inform Claude of the list of allowed tools when permission is denied
356- Fixed security vulnerability in Bash tool permission checks
357- Improved VSCode extension performance for large files
358
359## 1.0.123
360
361- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
362- Fixed thinking mode triggering on negation phrases like "don't think"
363- Fixed rendering performance degradation during token streaming
364- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
365- Enhanced BashTool environment snapshot logging
366- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
367- Migrated --debug logging to a file, to enable easy tailing & filtering
368
369## 1.0.120
370
371- Fix input lag during typing, especially noticeable with large prompts
372- Improved VSCode extension command registry and sessions dialog user experience
373- Enhanced sessions dialog responsiveness and visual feedback
374- Fixed IDE compatibility issue by removing worktree support check
375- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
376
377## 1.0.119
378
379- Fix Windows issue where process visually freezes on entering interactive mode
380- Support dynamic headers for MCP servers via headersHelper configuration
381- Fix thinking mode not working in headless sessions
382- Fix slash commands now properly update allowed tools instead of replacing them
383
384## 1.0.117
385
386- Add Ctrl-R history search to recall previous commands like bash/zsh
387- Fix input lag while typing, especially on Windows
388- Add sed command to auto-allowed commands in acceptEdits mode
389- Fix Windows PATH comparison to be case-insensitive for drive letters
390- Add permissions management hint to /add-dir output
391
392## 1.0.115
393
394- Improve thinking mode display with enhanced visual effects
395- Type /t to temporarily disable thinking mode in your prompt
396- Improve path validation for glob and grep tools
397- Show condensed output for post-tool hooks to reduce visual clutter
398- Fix visual feedback when loading state completes
399- Improve UI consistency for permission request dialogs
400
401## 1.0.113
402
403- Deprecated piped input in interactive mode
404- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
405
406## 1.0.112
407
408- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
409- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
410- Hooks: Added systemMessage support for SessionEnd hooks
411- Added `spinnerTipsEnabled` setting to disable spinner tips
412- IDE: Various improvements and bug fixes
413
414## 1.0.111
415
416- /model now validates provided model names
417- Fixed Bash tool crashes caused by malformed shell syntax parsing
418
419## 1.0.110
420
421- /terminal-setup command now supports WezTerm
422- MCP: OAuth tokens now proactively refresh before expiration
423- Fixed reliability issues with background Bash processes
424
425## 1.0.109
426
427- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
428
429## 1.0.106
430
431- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
432
433## 1.0.97
434
435- Settings: /doctor now validates permission rule syntax and suggests corrections
436
437## 1.0.94
438
439- Vertex: add support for global endpoints for supported models
440- /memory command now allows direct editing of all imported memory files
441- SDK: Add custom tools as callbacks
442- Added /todos command to list current todo items
443
444## 1.0.93
445
446- Windows: Add alt + v shortcut for pasting images from clipboard
447- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
448
449## 1.0.90
450
451- Settings file changes take effect immediately - no restart required
452
453## 1.0.88
454
455- Fixed issue causing "OAuth authentication is currently not supported"
456- Status line input now includes `exceeds_200k_tokens`
457- Fixed incorrect usage tracking in /cost.
458- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
459- Bedrock: Updated default Sonnet model to Sonnet 4
460
461## 1.0.86
462
463- Added /context to help users self-serve debug context issues
464- SDK: Added UUID support for all SDK messages
465- SDK: Added `--replay-user-messages` to replay user messages back to stdout
466
467## 1.0.85
468
469- Status line input now includes session cost info
470- Hooks: Introduced SessionEnd hook
471
472## 1.0.84
473
474- Fix tool_use/tool_result id mismatch error when network is unstable
475- Fix Claude sometimes ignoring real-time steering when wrapping up a task
476- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
477- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
478
479## 1.0.83
480
481- @-mention: Support files with spaces in path
482- New shimmering spinner
483
484## 1.0.82
485
486- SDK: Add request cancellation support
487- SDK: New additionalDirectories option to search custom paths, improved slash command processing
488- Settings: Validation prevents invalid fields in .claude/settings.json files
489- MCP: Improve tool name consistency
490- Bash: Fix crash when Claude tries to automatically read large files
491
492## 1.0.81
493
494- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
495- Agents: Fix custom agent loading when agent files are unparsable
496
497## 1.0.80
498
499- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
500
501## 1.0.77
502
503- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
504- SDK: Add session support and permission denial tracking
505- Fix token limit errors in conversation summarization
506- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
507
508## 1.0.73
509
510- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
511- MCP: Press Esc to cancel OAuth authentication flows
512- Bash: Improved command validation and reduced false security warnings
513- UI: Enhanced spinner animations and status line visual hierarchy
514- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
515
516## 1.0.72
517
518- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
519
520## 1.0.71
521
522- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
523- Customizable status line: add your terminal prompt to Claude Code with /statusline
524
525## 1.0.70
526
527- Performance: Optimized message rendering for better performance with large contexts
528- Windows: Fixed native file search, ripgrep, and subagent functionality
529- Added support for @-mentions in slash command arguments
530
531## 1.0.69
532
533- Upgraded Opus to version 4.1
534
535## 1.0.68
536
537- Fix incorrect model names being used for certain commands like `/pr-comments`
538- 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.
539- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
540- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
541- SDK: Added canUseTool callback support for tool confirmation
542- Added `disableAllHooks` setting
543- Improved file suggestions performance in large repos
544
545## 1.0.65
546
547- IDE: Fixed connection stability issues and error handling for diagnostics
548- Windows: Fixed shell environment setup for users without .bashrc files
549
550## 1.0.64
551
552- Agents: Added model customization support - you can now specify which model an agent should use
553- Agents: Fixed unintended access to the recursive agent tool
554- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
555- SDK: Fixed user input tracking across multi-turn conversations
556- Added hidden files to file search and @-mention suggestions
557
558## 1.0.63
559
560- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
561
562## 1.0.62
563
564- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
565- Hooks: Added SessionStart hook for new session initialization
566- /add-dir command now supports typeahead for directory paths
567- Improved network connectivity check reliability
568
569## 1.0.61
570
571- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
572- Settings: Added `--settings` flag to load settings from a JSON file
573- Settings: Fixed resolution of settings files paths that are symlinks
574- OTEL: Fixed reporting of wrong organization after authentication changes
575- Slash commands: Fixed permissions checking for allowed-tools with Bash
576- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
577- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
578- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
579
580## 1.0.60
581
582- You can now create custom subagents for specialized tasks! Run /agents to get started
583
584## 1.0.59
585
586- SDK: Added tool confirmation support with canUseTool callback
587- SDK: Allow specifying env for spawned process
588- Hooks: Exposed PermissionDecision to hooks (including "ask")
589- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
590- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
591
592## 1.0.58
593
594- Added support for reading PDFs
595- MCP: Improved server health status display in 'claude mcp list'
596- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
597
598## 1.0.57
599
600- Added support for specifying a model in slash commands
601- Improved permission messages to help Claude understand allowed tools
602- Fix: Remove trailing newlines from bash output in terminal wrapping
603
604## 1.0.56
605
606- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
607- Fixes for WSL IDE detection
608- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
609
610## 1.0.55
611
612- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
613- Windows: fixed Ctrl+Z crash
614- SDK: Added ability to capture error logging
615- Add --system-prompt-file option to override system prompt in print mode
616
617## 1.0.54
618
619- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
620- Custom slash commands: Added argument-hint to frontmatter
621- Windows: OAuth uses port 45454 and properly constructs browser URL
622- Windows: mode switching now uses alt + m, and plan mode renders properly
623- Shell: Switch to in-memory shell snapshot to fix file-related errors
624
625## 1.0.53
626
627- Updated @-mention file truncation from 100 lines to 2000 lines
628- 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).
629
630## 1.0.52
631
632- Added support for MCP server instructions
633
634## 1.0.51
635
636- Added support for native Windows (requires Git for Windows)
637- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
638- Settings: /doctor can now help you identify and fix invalid setting files
639- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
640- Increased auto-compact warning threshold from 60% to 80%
641- Fixed an issue with handling user directories with spaces for shell snapshots
642- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
643- Custom slash commands: Fixed user-level commands in subdirectories
644- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
645
646## 1.0.48
647
648- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
649- Added progress messages to Bash tool based on the last 5 lines of command output
650- Added expanding variables support for MCP server configuration
651- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
652- Improved IDE extension path handling when Claude Code runs in WSL
653- Hooks: Added a PreCompact hook
654- Vim mode: Added c, f/F, t/T
655
656## 1.0.45
657
658- Redesigned Search (Grep) tool with new tool input parameters and features
659- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
660- Fixed config file corruption issue by enforcing atomic writes
661- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
662- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
663- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
664
665## 1.0.44
666
667- New /export command lets you quickly export a conversation for sharing
668- MCP: resource_link tool results are now supported
669- MCP: tool annotations and tool titles now display in /mcp view
670- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
671
672## 1.0.43
673
674- Fixed a bug where the theme selector was saving excessively
675- Hooks: Added EPIPE system error handling
676
677## 1.0.42
678
679- Added tilde (`~`) expansion support to `/add-dir` command
680
681## 1.0.41
682
683- Hooks: Split Stop hook triggering into Stop and SubagentStop
684- Hooks: Enabled optional timeout configuration for each command
685- Hooks: Added "hook_event_name" to hook input
686- Fixed a bug where MCP tools would display twice in tool list
687- New tool parameters JSON for Bash tool in `tool_decision` event
688
689## 1.0.40
690
691- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
692
693## 1.0.39
694
695- New Active Time metric in OpenTelemetry logging
696
697## 1.0.38
698
699- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
700
701## 1.0.37
702
703- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
704
705## 1.0.36
706
707- Web search now takes today's date into context
708- Fixed a bug where stdio MCP servers were not terminating properly on exit
709
710## 1.0.35
711
712- Added support for MCP OAuth Authorization Server discovery
713
714## 1.0.34
715
716- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
717
718## 1.0.33
719
720- Improved logging functionality with session ID support
721- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
722- Improvements to plan mode
723
724## 1.0.32
725
726- Updated loopback config for litellm
727- Added forceLoginMethod setting to bypass login selection screen
728
729## 1.0.31
730
731- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
732
733## 1.0.30
734
735- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
736- Improved file path autocomplete with filename matching
737- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
738- Enhanced jq regex support for complex filters with pipes and select
739
740## 1.0.29
741
742- Improved CJK character support in cursor navigation and rendering
743
744## 1.0.28
745
746- Slash commands: Fix selector display during history navigation
747- Resizes images before upload to prevent API size limit errors
748- Added XDG_CONFIG_HOME support to configuration directory
749- Performance optimizations for memory usage
750- New attributes (terminal.type, language) in OpenTelemetry logging
751
752## 1.0.27
753
754- Streamable HTTP MCP servers are now supported
755- Remote MCP servers (SSE and HTTP) now support OAuth
756- MCP resources can now be @-mentioned
757- /resume slash command to switch conversations within Claude Code
758
759## 1.0.25
760
761- Slash commands: moved "project" and "user" prefixes to descriptions
762- Slash commands: improved reliability for command discovery
763- Improved support for Ghostty
764- Improved web search reliability
765
766## 1.0.24
767
768- Improved /mcp output
769- Fixed a bug where settings arrays got overwritten instead of merged
770
771## 1.0.23
772
773- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
774- Released Python SDK: pip install claude-code-sdk to get started
775
776## 1.0.22
777
778- SDK: Renamed `total_cost` to `total_cost_usd`
779
780## 1.0.21
781
782- Improved editing of files with tab-based indentation
783- Fix for tool_use without matching tool_result errors
784- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
785
786## 1.0.18
787
788- Added --add-dir CLI argument for specifying additional working directories
789- Added streaming input support without require -p flag
790- Improved startup performance and session storage performance
791- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
792- Added detailed MCP server tools display (/mcp)
793- MCP authentication and permission improvements
794- Added auto-reconnection for MCP SSE connections on disconnect
795- Fixed issue where pasted content was lost when dialogs appeared
796
797## 1.0.17
798
799- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
800- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
801- MCP server list UI improvements
802- Update Claude Code process title to display "claude" instead of "node"
803
804## 1.0.11
805
806- Claude Code can now also be used with a Claude Pro subscription
807- Added /upgrade for smoother switching to Claude Max plans
808- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
809- Improved shell configuration error handling
810- Improved todo list handling during compaction
811
812## 1.0.10
813
814- Added markdown table support
815- Improved streaming performance
816
817## 1.0.8
818
819- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
820- Increased default otel interval from 1s -> 5s
821- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
822- Fixed a regression where search tools unnecessarily asked for permissions
823- Added support for triggering thinking non-English languages
824- Improved compacting UI
825
826## 1.0.7
827
828- Renamed /allowed-tools -> /permissions
829- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
830- Deprecated claude config commands in favor of editing settings.json
831- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
832- Improved error handling for /install-github-app
833- Bugfixes, UI polish, and tool reliability improvements
834
835## 1.0.6
836
837- Improved edit reliability for tab-indented files
838- Respect CLAUDE_CONFIG_DIR everywhere
839- Reduced unnecessary tool permission prompts
840- Added support for symlinks in @file typeahead
841- Bugfixes, UI polish, and tool reliability improvements
842
843## 1.0.4
844
845- Fixed a bug where MCP tool errors weren't being parsed correctly
846
847## 1.0.1
848
849- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
850- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
851- Updated documentation links and OAuth process descriptions
852
853## 1.0.0
854
855- Claude Code is now generally available
856- Introducing Sonnet 4 and Opus 4 models
857
858## 0.2.125
859
860- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
861- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
862
863## 0.2.117
864
865- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
866- Introduced settings.cleanupPeriodDays
867- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
868- Introduced --debug mode
869
870## 0.2.108
871
872- You can now send messages to Claude while it works to steer Claude in real-time
873- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
874- Fixed a bug where thinking was not working in -p mode
875- Fixed a regression in /cost reporting
876- Deprecated MCP wizard interface in favor of other MCP commands
877- Lots of other bugfixes and improvements
878
879## 0.2.107
880
881- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
882
883## 0.2.106
884
885- MCP SSE server configs can now specify custom headers
886- Fixed a bug where MCP permission prompt didn't always show correctly
887
888## 0.2.105
889
890- Claude can now search the web
891- Moved system & account status to /status
892- Added word movement keybindings for Vim
893- Improved latency for startup, todo tool, and file edits
894
895## 0.2.102
896
897- Improved thinking triggering reliability
898- Improved @mention reliability for images and folders
899- You can now paste multiple large chunks into one prompt
900
901## 0.2.100
902
903- Fixed a crash caused by a stack overflow error
904- Made db storage optional; missing db support disables --continue and --resume
905
906## 0.2.98
907
908- Fixed an issue where auto-compact was running twice
909
910## 0.2.96
911
912- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
913
914## 0.2.93
915
916- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
917- Claude now has access to a Todo list that helps it stay on track and be more organized
918
919## 0.2.82
920
921- Added support for --disallowedTools
922- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
923
924## 0.2.75
925
926- Hit Enter to queue up additional messages while Claude is working
927- Drag in or copy/paste image files directly into the prompt
928- @-mention files to directly add them to context
929- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
930- Improved performance for filename auto-complete
931
932## 0.2.74
933
934- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
935- Task tool can now perform writes and run bash commands
936
937## 0.2.72
938
939- Updated spinner to indicate tokens loaded and tool usage
940
941## 0.2.70
942
943- Network commands like curl are now available for Claude to use
944- Claude can now run multiple web queries in parallel
945- Pressing ESC once immediately interrupts Claude in Auto-accept mode
946
947## 0.2.69
948
949- Fixed UI glitches with improved Select component behavior
950- Enhanced terminal output display with better text truncation logic
951
952## 0.2.67
953
954- Shared project permission rules can be saved in .claude/settings.json
955
956## 0.2.66
957
958- Print mode (-p) now supports streaming output via --output-format=stream-json
959- Fixed issue where pasting could trigger memory or bash mode unexpectedly
960
961## 0.2.63
962
963- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
964
965## 0.2.61
966
967- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
968- Enhanced image detection for more reliable clipboard paste functionality
969- Fixed an issue where ESC key could crash the conversation history selector
970
971## 0.2.59
972
973- Copy+paste images directly into your prompt
974- Improved progress indicators for bash and fetch tools
975- Bugfixes for non-interactive mode (-p)
976
977## 0.2.54
978
979- Quickly add to Memory by starting your message with '#'
980- Press ctrl+r to see full output for long tool results
981- Added support for MCP SSE transport
982
983## 0.2.53
984
985- New web fetch tool lets Claude view URLs that you paste in
986- Fixed a bug with JPEG detection
987
988## 0.2.50
989
990- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
991
992## 0.2.49
993
994- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
995
996## 0.2.47
997
998- Press Tab to auto-complete file and folder names
999- Press Shift + Tab to toggle auto-accept for file edits
1000- Automatic conversation compaction for infinite conversation length (toggle with /config)
1001
1002## 0.2.44
1003
1004- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1005
1006## 0.2.41
1007
1008- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1009- MCP server startup no longer blocks the app from starting up
1010
1011## 0.2.37
1012
1013- New /release-notes command lets you view release notes at any time
1014- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1015
1016## 0.2.36
1017
1018- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1019- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1020
1021## 0.2.34
1022
1023- Vim bindings for text input - enable with /vim or /config
1024
1025## 0.2.32
1026
1027- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1028- Fix for some PersistentShell issues
1029
1030## 0.2.31
1031
1032- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1033- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1034
1035## 0.2.30
1036
1037- Added ANSI color theme for better terminal compatibility
1038- Fixed issue where slash command arguments weren't being sent properly
1039- (Mac-only) API keys are now stored in macOS Keychain
1040
1041## 0.2.26
1042
1043- New /approved-tools command for managing tool permissions
1044- Word-level diff display for improved code readability
1045- Fuzzy matching for slash commands
1046
1047## 0.2.21
1048
1049- Fuzzy matching for /commands