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 553d6ff · on Jan 9, 2026 History
1,323 lines · 62.3 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.2
4
5- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
6- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
7- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
8- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
9- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
10- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
11- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
12- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
13- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
14- Fixed updates incorrectly claiming another installation is in progress
15- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
16- Fixed remote session URL and teleport being broken when using `/tasks` command
17- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
18- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
19- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
20- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
21- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
22- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
23- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
24- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
25- [SDK] Changed minimum zod peer dependency to ^4.0.0
26- [VSCode] Fixed usage display not updating after manual compact
27
28## 2.1.0
29
30- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
31- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
32- Added support for `agent` field in skills to specify agent type for execution
33- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
34- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
35- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
36- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
37- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
38- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
39- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
40- Fixed Esc key with queued prompts to only move them to input without canceling the running task
41- Reduced permission prompts for complex bash commands
42- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
43- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
44- Fixed plugin path resolution for file-based marketplace sources
45- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
46- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
47- Fixed Claude in Chrome support for WSL environments
48- Fixed Windows native installer silently failing when executable creation fails
49- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
50- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
51- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
52- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
53- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
54- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
55- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
56- Added hooks support for skill and slash command frontmatter
57- Added new Vim motions: `;` and `,` to repeat f/F/t/T motions, `y` operator for yank with `yy`/`Y`, `p`/`P` for paste, text objects (`iw`, `aw`, `iW`, `aW`, `i"`, `a"`, `i'`, `a'`, `i(`, `a(`, `i[`, `a[`, `i{`, `a{`), `>>` and `<<` for indent/dedent, and `J` to join lines
58- Added `/plan` command shortcut to enable plan mode directly from the prompt
59- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
60- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
61- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
62- Added support for `once: true` config for hooks
63- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
64- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
65- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
66- Added left/right arrow key navigation for cycling through tabs in dialogs
67- Added real-time thinking block display in Ctrl+O transcript mode
68- Added filepath to full output in background bash task details dialog
69- Added Skills as a separate category in the context visualization
70- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
71- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
72- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
73- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
74- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
75- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
76- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
77- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
78- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
79- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
80- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
81- Fixed commands with `$()` command substitution failing with parse errors
82- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
83- Fixed bash command prefix extraction to correctly identify subcommands after global options (e.g., `git -C /path log` now correctly matches `Bash(git log:*)` rules)
84- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
85- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
86- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
87- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
88- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
89- Fixed subagents sometimes not inheriting the parent's model by default
90- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
91- Fixed duplicate Bash commands appearing in permission request option labels
92- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
93- Fixed background task completion notifications to appear proactively with bullet point
94- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
95- Fixed cursor disappearing after dismissing permission dialogs
96- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
97- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
98- Fixed images being silently dropped when queueing messages while backgrounding a task
99- Fixed large pasted images failing with "Image was too large" error
100- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
101- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
102- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
103- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
104- Fixed spinner token counter to properly accumulate tokens from subagents during execution
105- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
106- Fixed race condition where LSP tool could return "no server available" during startup
107- Fixed feedback submission hanging indefinitely when network requests timeout
108- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
109- Fixed hook success message showing trailing colon when hook has no output
110- Multiple optimizations to improve startup performance
111- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
112- Improved performance when reading Jupyter notebooks with many cells
113- Improved reliability for piped input like `cat refactor.md | claude`
114- Improved reliability for AskQuestion tool
115- Improved sed in-place edit commands to render as file edits with diff preview
116- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
117- Improved compaction reliability
118- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
119- Improved skills to show progress while executing, displaying tool uses as they happen
120- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
121- Improved skill suggestions to prioritize recently and frequently used skills
122- Improved spinner feedback when waiting for the first response token
123- Improved token count display in spinner to include tokens from background agents
124- Improved incremental output for async agents to give the main thread more control and visibility
125- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
126- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
127- Improved macOS screenshot paste reliability with TIFF format support
128- Improved `/stats` output
129- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
130- Changed "Interrupted" message color from red to grey for a less alarming appearance
131- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
132- Removed underline styling from image reference links
133- [SDK] Changed minimum zod peer dependency to ^4.0.0
134- [VSCode] Added currently selected model name to the context menu
135- [VSCode] Added descriptive labels on auto-accept permission button (e.g., "Yes, allow npm for this project" instead of "Yes, and don't ask again")
136- [VSCode] Fixed paragraph breaks not rendering in markdown content
137- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
138- [Windows] Fixed issue with improper rendering
139
140## 2.0.76
141
142- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
143
144## 2.0.75
145
146- Minor bugfixes
147
148## 2.0.74
149
150- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
151- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
152- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
153- Added syntax highlighting info to theme picker
154- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
155- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
156- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
157- Fixed a potential crash when syntax highlighting isn't initialized correctly
158- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
159- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
160- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
161- [Windows] Fixed issue with improper rendering
162- [VSCode] Added gift tag pictogram for year-end promotion message
163
164## 2.0.73
165
166- Added clickable `[Image #N]` links that open attached images in the default viewer
167- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
168- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
169- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
170- Fixed slow input history cycling and race condition that could overwrite text after message submission
171- Improved `/theme` command to open theme picker directly
172- Improved theme picker UI
173- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
174- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
175
176## 2.0.72
177
178- Added Claude in Chrome (Beta) feature that works with the Chrome extension (https://claude.ai/chrome) to let you control your browser directly from Claude Code
179- Reduced terminal flickering
180- Added scannable QR code to mobile app tip for quick app downloads
181- Added loading indicator when resuming conversations for better feedback
182- Fixed `/context` command not respecting custom system prompts in non-interactive mode
183- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
184- Improved @ mention file suggestion speed (~3x faster in git repositories)
185- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
186- Improved settings validation errors to be more prominent
187- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
188
189## 2.0.71
190
191- Added /config toggle to enable/disable prompt suggestions
192- Added `/settings` as an alias for the `/config` command
193- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
194- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
195- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
196- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
197- New syntax highlighting engine for native build
198
199## 2.0.70
200
201- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
202- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
203- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
204- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
205- Fixed input being cleared when processing queued commands while the user was typing
206- Fixed prompt suggestions replacing typed input when pressing Tab
207- Fixed diff view not updating when terminal is resized
208- Improved memory usage by 3x for large conversations
209- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
210- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
211- Fix thinking mode toggle in /config not persisting correctly
212- Improve UI for file creation permission dialog
213
214## 2.0.69
215
216- Minor bugfixes
217
218## 2.0.68
219
220- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
221- Fixed a bug where disallowed MCP tools were visible to the model
222- Fixed an issue where steering messages could be lost while a subagent is working
223- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
224- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
225- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
226
227## 2.0.67
228
229- Thinking mode is now enabled by default for Opus 4.5
230- Thinking mode configuration has moved to /config
231- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
232- Show reason why autoupdater is disabled in `/doctor`
233- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
234- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
235- Fixed scroll position resetting after deleting a permission rule in `/permissions`
236- 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
237- Fixed `claude install --force` not bypassing stale lock files
238- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
239- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
240
241## 2.0.65
242
243- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
244- Added context window information to status line input
245- Added `fileSuggestion` setting for custom `@` file search commands
246- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
247- Fixed prompt not being saved to history when aborting a query with Escape
248- Fixed Read tool image handling to identify format from bytes instead of file extension
249
250## 2.0.64
251
252- Made auto-compacting instant
253- Agents and bash commands can run asynchronously and send messages to wake up the main agent
254- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
255- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
256- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
257- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
258- Fixed auto-loading .env when using native installer
259- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
260- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
261- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
262- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
263- Bedrock: Improve efficiency of token counting
264- Bedrock: Add support for `aws login` AWS Management Console credentials
265- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
266
267## 2.0.62
268
269- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
270- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
271- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
272- Fixed slash command selection not working when multiple commands share the same name
273- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
274- Fixed running versions getting removed because lock file incorrectly going stale
275- Fixed IDE diff tab not closing when rejecting file changes
276
277## 2.0.61
278
279- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
280
281## 2.0.60
282
283- Added background agent support. Agents run in the background while you work
284- Added --disable-slash-commands CLI flag to disable all slash commands
285- Added model name to "Co-Authored-By" commit messages
286- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
287- Updated Fetch to skip summarization for pre-approved websites
288- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
289
290## 2.0.59
291
292- Added --agent CLI flag to override the agent setting for the current session
293- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
294- VS Code: Fixed .claude.json config file being read from incorrect location
295
296## 2.0.58
297
298- Pro users now have access to Opus 4.5 as part of their subscription!
299- Fixed timer duration showing "11m 60s" instead of "12m 0s"
300- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
301
302## 2.0.57
303
304- Added feedback input when rejecting plans, allowing users to tell Claude what to change
305- VSCode: Added streaming message support for real-time response display
306
307## 2.0.56
308
309- Added setting to enable/disable terminal progress bar (OSC 9;4)
310- 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.
311
312## 2.0.55
313
314- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
315- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
316- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
317- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
318
319## 2.0.54
320
321- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
322- Fix issue with excessive iTerm notifications
323
324## 2.0.52
325
326- Fixed duplicate message display when starting Claude with a command line argument
327- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
328- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
329- Permit some uses of `$!` in bash commands
330
331## 2.0.51
332
333- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
334- Introducing Claude Code for Desktop: https://claude.com/download
335- 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
336- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
337- Plan Mode now builds more precise plans and executes more thoroughly
338- Usage limit notifications now easier to understand
339- Switched `/usage` back to "% used"
340- Fixed handling of thinking errors
341- Fixed performance regression
342
343## 2.0.50
344
345- Fixed bug preventing calling MCP tools that have nested references in their input schemas
346- Silenced a noisy but harmless error during upgrades
347- Improved ultrathink text display
348- Improved clarity of 5-hour session limit warning message
349
350## 2.0.49
351
352- Added readline-style ctrl-y for pasting deleted text
353- Improved clarity of usage limit warning message
354- Fixed handling of subagent permissions
355
356## 2.0.47
357
358- Improved error messages and validation for `claude --teleport`
359- Improved error handling in `/usage`
360- Fixed race condition with history entry not getting logged at exit
361- Fixed Vertex AI configuration not being applied from `settings.json`
362
363## 2.0.46
364
365- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
366
367## 2.0.45
368
369- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
370- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
371- Send background tasks to Claude Code on the web by starting a message with `&`
372
373## 2.0.43
374
375- Added `permissionMode` field for custom agents
376- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
377- Added skills frontmatter field to declare skills to auto-load for subagents
378- Added the `SubagentStart` hook event
379- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
380- Fixed duplicate rendering of some messages in the UI
381- Fixed some visual flickers
382- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
383
384## 2.0.42
385
386- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
387
388## 2.0.41
389
390- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
391- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
392- Fixed incorrect labeling of user settings vs project settings in command descriptions
393- Fixed crash when plugin command hooks timeout during execution
394- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
395- Fixed broken security documentation links in trust dialogs and onboarding
396- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
397- ctrl-r history search landing on a slash command no longer cancels the search
398- SDK: Support custom timeouts for hooks
399- Allow more safe git commands to run without approval
400- Plugins: Added support for sharing and installing output styles
401- Teleporting a session from web will automatically set the upstream branch
402
403## 2.0.37
404
405- Fixed how idleness is computed for notifications
406- Hooks: Added matcher values for Notification hook events
407- Output Styles: Added `keep-coding-instructions` option to frontmatter
408
409## 2.0.36
410
411- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
412- Fixed queued messages being incorrectly executed as bash commands
413- Fixed input being lost when typing while a queued message is processed
414
415## 2.0.35
416
417- Improve fuzzy search results when searching commands
418- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
419- 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
420- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
421- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
422
423## 2.0.34
424
425- VSCode Extension: Added setting to configure the initial permission mode for new conversations
426- Improved file path suggestion performance with native Rust-based fuzzy finder
427- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
428- Fixed memory crash when reading or writing large files (especially base64-encoded images)
429
430## 2.0.33
431
432- Native binary installs now launch quicker.
433- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
434- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
435
436## 2.0.32
437
438- Un-deprecate output styles based on community feedback
439- Added `companyAnnouncements` setting for displaying announcements on startup
440- Fixed hook progress messages not updating correctly during PostToolUse hook execution
441
442## 2.0.31
443
444- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
445- Vertex: add support for Web Search on supported models
446- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
447- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
448- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
449- Fixed plugin uninstall not removing plugins
450
451## 2.0.30
452
453- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
454- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
455- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
456- Added prompt-based stop hooks
457- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
458- Enabled SSE MCP servers on native build
459- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
460- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
461- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
462- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
463- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
464- Fixed bug that saved session permissions to local settings
465- Fixed MCP tools not being available to sub-agents
466- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
467- Fixed delay when navigating through typeahead suggestions with arrow keys
468- VSCode: Restored selection indicator in input footer showing current file or code selection status
469
470## 2.0.28
471
472- Plan mode: introduced new Plan subagent
473- Subagents: claude can now choose to resume subagents
474- Subagents: claude can dynamically choose the model used by its subagents
475- SDK: added --max-budget-usd flag
476- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
477- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
478- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
479- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
480- Various other bug fixes
481
482## 2.0.27
483
484- New UI for permission prompts
485- Added current branch filtering and search to session resume screen for easier navigation
486- Fixed directory @-mention causing "No assistant message found" error
487- VSCode Extension: Add config setting to include .gitignored files in file searches
488- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
489
490## 2.0.25
491
492- 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
493
494## 2.0.24
495
496- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
497- Claude Code Web: Support for Web -> CLI teleport
498- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
499- Bedrock: Display awsAuthRefresh output when auth is required
500
501## 2.0.22
502
503- Fixed content layout shift when scrolling through slash commands
504- IDE: Add toggle to enable/disable thinking.
505- Fix bug causing duplicate permission prompts with parallel tool calls
506- Add support for enterprise managed MCP allowlist and denylist
507
508## 2.0.21
509
510- Support MCP `structuredContent` field in tool responses
511- Added an interactive question tool
512- Claude will now ask you questions more often in plan mode
513- Added Haiku 4.5 as a model option for Pro users
514- Fixed an issue where queued commands don't have access to previous messages' output
515
516## 2.0.20
517
518- Added support for Claude Skills
519
520## 2.0.19
521
522- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
523- Fixed a bug where Haiku was unnecessarily called in print mode
524
525## 2.0.17
526
527- Added Haiku 4.5 to model selector!
528- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
529- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
530- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
531- OTEL: support HTTP_PROXY and HTTPS_PROXY
532- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
533
534## 2.0.15
535
536- Fixed bug with resuming where previously created files needed to be read again before writing
537- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
538
539## 2.0.14
540
541- Fix @-mentioning MCP servers to toggle them on/off
542- Improve permission checks for bash with inline env vars
543- Fix ultrathink + thinking toggle
544- Reduce unnecessary logins
545- Document --system-prompt
546- Several improvements to rendering
547- Plugins UI polish
548
549## 2.0.13
550
551- Fixed `/plugin` not working on native build
552
553## 2.0.12
554
555- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
556- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
557- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
558- `/plugin validate` command for validating plugin structure and configuration
559- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
560- Plugin documentation available at https://code.claude.com/docs/en/plugins
561- Comprehensive error messages and diagnostics via `/doctor` command
562- Avoid flickering in `/model` selector
563- Improvements to `/help`
564- Avoid mentioning hooks in `/resume` summaries
565- Changes to the "verbose" setting in `/config` now persist across sessions
566
567## 2.0.11
568
569- Reduced system prompt size by 1.4k tokens
570- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
571- Fixed Opus fallback rate limit errors appearing incorrectly
572- Fixed /add-dir command selecting wrong default tab
573
574## 2.0.10
575
576- Rewrote terminal renderer for buttery smooth UI
577- Enable/disable MCP servers by @mentioning, or in /mcp
578- Added tab completion for shell commands in bash mode
579- PreToolUse hooks can now modify tool inputs
580- Press Ctrl-G to edit your prompt in your system's configured text editor
581- Fixes for bash permission checks with environment variables in the command
582
583## 2.0.9
584
585- Fix regression where bash backgrounding stopped working
586
587## 2.0.8
588
589- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
590- IDE: Add drag-and-drop support for files and folders in chat
591- /context: Fix counting for thinking blocks
592- Improve message rendering for users with light themes on dark terminals
593- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
594
595## 2.0.5
596
597- IDE: Fix IME unintended message submission with Enter and Tab
598- IDE: Add "Open in Terminal" link in login screen
599- Fix unhandled OAuth expiration 401 API errors
600- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
601
602## 2.0.1
603
604- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
605- Various bug fixes and presentation improvements
606
607## 2.0.0
608
609- New native VS Code extension
610- Fresh coat of paint throughout the whole app
611- /rewind a conversation to undo code changes
612- /usage command to see plan limits
613- Tab to toggle thinking (sticky across sessions)
614- Ctrl-R to search history
615- Unshipped claude config command
616- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
617- SDK: The Claude Code SDK is now the Claude Agent SDK
618- Add subagents dynamically with `--agents` flag
619
620## 1.0.126
621
622- Enable /context command for Bedrock and Vertex
623- Add mTLS support for HTTP-based OpenTelemetry exporters
624
625## 1.0.124
626
627- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
628- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
629- No longer inform Claude of the list of allowed tools when permission is denied
630- Fixed security vulnerability in Bash tool permission checks
631- Improved VSCode extension performance for large files
632
633## 1.0.123
634
635- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
636- Fixed thinking mode triggering on negation phrases like "don't think"
637- Fixed rendering performance degradation during token streaming
638- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
639- Enhanced BashTool environment snapshot logging
640- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
641- Migrated --debug logging to a file, to enable easy tailing & filtering
642
643## 1.0.120
644
645- Fix input lag during typing, especially noticeable with large prompts
646- Improved VSCode extension command registry and sessions dialog user experience
647- Enhanced sessions dialog responsiveness and visual feedback
648- Fixed IDE compatibility issue by removing worktree support check
649- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
650
651## 1.0.119
652
653- Fix Windows issue where process visually freezes on entering interactive mode
654- Support dynamic headers for MCP servers via headersHelper configuration
655- Fix thinking mode not working in headless sessions
656- Fix slash commands now properly update allowed tools instead of replacing them
657
658## 1.0.117
659
660- Add Ctrl-R history search to recall previous commands like bash/zsh
661- Fix input lag while typing, especially on Windows
662- Add sed command to auto-allowed commands in acceptEdits mode
663- Fix Windows PATH comparison to be case-insensitive for drive letters
664- Add permissions management hint to /add-dir output
665
666## 1.0.115
667
668- Improve thinking mode display with enhanced visual effects
669- Type /t to temporarily disable thinking mode in your prompt
670- Improve path validation for glob and grep tools
671- Show condensed output for post-tool hooks to reduce visual clutter
672- Fix visual feedback when loading state completes
673- Improve UI consistency for permission request dialogs
674
675## 1.0.113
676
677- Deprecated piped input in interactive mode
678- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
679
680## 1.0.112
681
682- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
683- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
684- Hooks: Added systemMessage support for SessionEnd hooks
685- Added `spinnerTipsEnabled` setting to disable spinner tips
686- IDE: Various improvements and bug fixes
687
688## 1.0.111
689
690- /model now validates provided model names
691- Fixed Bash tool crashes caused by malformed shell syntax parsing
692
693## 1.0.110
694
695- /terminal-setup command now supports WezTerm
696- MCP: OAuth tokens now proactively refresh before expiration
697- Fixed reliability issues with background Bash processes
698
699## 1.0.109
700
701- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
702
703## 1.0.106
704
705- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
706
707## 1.0.97
708
709- Settings: /doctor now validates permission rule syntax and suggests corrections
710
711## 1.0.94
712
713- Vertex: add support for global endpoints for supported models
714- /memory command now allows direct editing of all imported memory files
715- SDK: Add custom tools as callbacks
716- Added /todos command to list current todo items
717
718## 1.0.93
719
720- Windows: Add alt + v shortcut for pasting images from clipboard
721- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
722
723## 1.0.90
724
725- Settings file changes take effect immediately - no restart required
726
727## 1.0.88
728
729- Fixed issue causing "OAuth authentication is currently not supported"
730- Status line input now includes `exceeds_200k_tokens`
731- Fixed incorrect usage tracking in /cost.
732- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
733- Bedrock: Updated default Sonnet model to Sonnet 4
734
735## 1.0.86
736
737- Added /context to help users self-serve debug context issues
738- SDK: Added UUID support for all SDK messages
739- SDK: Added `--replay-user-messages` to replay user messages back to stdout
740
741## 1.0.85
742
743- Status line input now includes session cost info
744- Hooks: Introduced SessionEnd hook
745
746## 1.0.84
747
748- Fix tool_use/tool_result id mismatch error when network is unstable
749- Fix Claude sometimes ignoring real-time steering when wrapping up a task
750- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
751- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
752
753## 1.0.83
754
755- @-mention: Support files with spaces in path
756- New shimmering spinner
757
758## 1.0.82
759
760- SDK: Add request cancellation support
761- SDK: New additionalDirectories option to search custom paths, improved slash command processing
762- Settings: Validation prevents invalid fields in .claude/settings.json files
763- MCP: Improve tool name consistency
764- Bash: Fix crash when Claude tries to automatically read large files
765
766## 1.0.81
767
768- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
769- Agents: Fix custom agent loading when agent files are unparsable
770
771## 1.0.80
772
773- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
774
775## 1.0.77
776
777- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
778- SDK: Add session support and permission denial tracking
779- Fix token limit errors in conversation summarization
780- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
781
782## 1.0.73
783
784- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
785- MCP: Press Esc to cancel OAuth authentication flows
786- Bash: Improved command validation and reduced false security warnings
787- UI: Enhanced spinner animations and status line visual hierarchy
788- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
789
790## 1.0.72
791
792- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
793
794## 1.0.71
795
796- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
797- Customizable status line: add your terminal prompt to Claude Code with /statusline
798
799## 1.0.70
800
801- Performance: Optimized message rendering for better performance with large contexts
802- Windows: Fixed native file search, ripgrep, and subagent functionality
803- Added support for @-mentions in slash command arguments
804
805## 1.0.69
806
807- Upgraded Opus to version 4.1
808
809## 1.0.68
810
811- Fix incorrect model names being used for certain commands like `/pr-comments`
812- 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.
813- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
814- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
815- SDK: Added canUseTool callback support for tool confirmation
816- Added `disableAllHooks` setting
817- Improved file suggestions performance in large repos
818
819## 1.0.65
820
821- IDE: Fixed connection stability issues and error handling for diagnostics
822- Windows: Fixed shell environment setup for users without .bashrc files
823
824## 1.0.64
825
826- Agents: Added model customization support - you can now specify which model an agent should use
827- Agents: Fixed unintended access to the recursive agent tool
828- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
829- SDK: Fixed user input tracking across multi-turn conversations
830- Added hidden files to file search and @-mention suggestions
831
832## 1.0.63
833
834- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
835
836## 1.0.62
837
838- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
839- Hooks: Added SessionStart hook for new session initialization
840- /add-dir command now supports typeahead for directory paths
841- Improved network connectivity check reliability
842
843## 1.0.61
844
845- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
846- Settings: Added `--settings` flag to load settings from a JSON file
847- Settings: Fixed resolution of settings files paths that are symlinks
848- OTEL: Fixed reporting of wrong organization after authentication changes
849- Slash commands: Fixed permissions checking for allowed-tools with Bash
850- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
851- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
852- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
853
854## 1.0.60
855
856- You can now create custom subagents for specialized tasks! Run /agents to get started
857
858## 1.0.59
859
860- SDK: Added tool confirmation support with canUseTool callback
861- SDK: Allow specifying env for spawned process
862- Hooks: Exposed PermissionDecision to hooks (including "ask")
863- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
864- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
865
866## 1.0.58
867
868- Added support for reading PDFs
869- MCP: Improved server health status display in 'claude mcp list'
870- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
871
872## 1.0.57
873
874- Added support for specifying a model in slash commands
875- Improved permission messages to help Claude understand allowed tools
876- Fix: Remove trailing newlines from bash output in terminal wrapping
877
878## 1.0.56
879
880- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
881- Fixes for WSL IDE detection
882- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
883
884## 1.0.55
885
886- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
887- Windows: fixed Ctrl+Z crash
888- SDK: Added ability to capture error logging
889- Add --system-prompt-file option to override system prompt in print mode
890
891## 1.0.54
892
893- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
894- Custom slash commands: Added argument-hint to frontmatter
895- Windows: OAuth uses port 45454 and properly constructs browser URL
896- Windows: mode switching now uses alt + m, and plan mode renders properly
897- Shell: Switch to in-memory shell snapshot to fix file-related errors
898
899## 1.0.53
900
901- Updated @-mention file truncation from 100 lines to 2000 lines
902- 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).
903
904## 1.0.52
905
906- Added support for MCP server instructions
907
908## 1.0.51
909
910- Added support for native Windows (requires Git for Windows)
911- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
912- Settings: /doctor can now help you identify and fix invalid setting files
913- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
914- Increased auto-compact warning threshold from 60% to 80%
915- Fixed an issue with handling user directories with spaces for shell snapshots
916- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
917- Custom slash commands: Fixed user-level commands in subdirectories
918- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
919
920## 1.0.48
921
922- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
923- Added progress messages to Bash tool based on the last 5 lines of command output
924- Added expanding variables support for MCP server configuration
925- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
926- Improved IDE extension path handling when Claude Code runs in WSL
927- Hooks: Added a PreCompact hook
928- Vim mode: Added c, f/F, t/T
929
930## 1.0.45
931
932- Redesigned Search (Grep) tool with new tool input parameters and features
933- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
934- Fixed config file corruption issue by enforcing atomic writes
935- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
936- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
937- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
938
939## 1.0.44
940
941- New /export command lets you quickly export a conversation for sharing
942- MCP: resource_link tool results are now supported
943- MCP: tool annotations and tool titles now display in /mcp view
944- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
945
946## 1.0.43
947
948- Fixed a bug where the theme selector was saving excessively
949- Hooks: Added EPIPE system error handling
950
951## 1.0.42
952
953- Added tilde (`~`) expansion support to `/add-dir` command
954
955## 1.0.41
956
957- Hooks: Split Stop hook triggering into Stop and SubagentStop
958- Hooks: Enabled optional timeout configuration for each command
959- Hooks: Added "hook_event_name" to hook input
960- Fixed a bug where MCP tools would display twice in tool list
961- New tool parameters JSON for Bash tool in `tool_decision` event
962
963## 1.0.40
964
965- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
966
967## 1.0.39
968
969- New Active Time metric in OpenTelemetry logging
970
971## 1.0.38
972
973- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
974
975## 1.0.37
976
977- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
978
979## 1.0.36
980
981- Web search now takes today's date into context
982- Fixed a bug where stdio MCP servers were not terminating properly on exit
983
984## 1.0.35
985
986- Added support for MCP OAuth Authorization Server discovery
987
988## 1.0.34
989
990- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
991
992## 1.0.33
993
994- Improved logging functionality with session ID support
995- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
996- Improvements to plan mode
997
998## 1.0.32
999
1000- Updated loopback config for litellm
1001- Added forceLoginMethod setting to bypass login selection screen
1002
1003## 1.0.31
1004
1005- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1006
1007## 1.0.30
1008
1009- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1010- Improved file path autocomplete with filename matching
1011- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1012- Enhanced jq regex support for complex filters with pipes and select
1013
1014## 1.0.29
1015
1016- Improved CJK character support in cursor navigation and rendering
1017
1018## 1.0.28
1019
1020- Slash commands: Fix selector display during history navigation
1021- Resizes images before upload to prevent API size limit errors
1022- Added XDG_CONFIG_HOME support to configuration directory
1023- Performance optimizations for memory usage
1024- New attributes (terminal.type, language) in OpenTelemetry logging
1025
1026## 1.0.27
1027
1028- Streamable HTTP MCP servers are now supported
1029- Remote MCP servers (SSE and HTTP) now support OAuth
1030- MCP resources can now be @-mentioned
1031- /resume slash command to switch conversations within Claude Code
1032
1033## 1.0.25
1034
1035- Slash commands: moved "project" and "user" prefixes to descriptions
1036- Slash commands: improved reliability for command discovery
1037- Improved support for Ghostty
1038- Improved web search reliability
1039
1040## 1.0.24
1041
1042- Improved /mcp output
1043- Fixed a bug where settings arrays got overwritten instead of merged
1044
1045## 1.0.23
1046
1047- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1048- Released Python SDK: pip install claude-code-sdk to get started
1049
1050## 1.0.22
1051
1052- SDK: Renamed `total_cost` to `total_cost_usd`
1053
1054## 1.0.21
1055
1056- Improved editing of files with tab-based indentation
1057- Fix for tool_use without matching tool_result errors
1058- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1059
1060## 1.0.18
1061
1062- Added --add-dir CLI argument for specifying additional working directories
1063- Added streaming input support without require -p flag
1064- Improved startup performance and session storage performance
1065- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1066- Added detailed MCP server tools display (/mcp)
1067- MCP authentication and permission improvements
1068- Added auto-reconnection for MCP SSE connections on disconnect
1069- Fixed issue where pasted content was lost when dialogs appeared
1070
1071## 1.0.17
1072
1073- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1074- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1075- MCP server list UI improvements
1076- Update Claude Code process title to display "claude" instead of "node"
1077
1078## 1.0.11
1079
1080- Claude Code can now also be used with a Claude Pro subscription
1081- Added /upgrade for smoother switching to Claude Max plans
1082- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1083- Improved shell configuration error handling
1084- Improved todo list handling during compaction
1085
1086## 1.0.10
1087
1088- Added markdown table support
1089- Improved streaming performance
1090
1091## 1.0.8
1092
1093- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1094- Increased default otel interval from 1s -> 5s
1095- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1096- Fixed a regression where search tools unnecessarily asked for permissions
1097- Added support for triggering thinking non-English languages
1098- Improved compacting UI
1099
1100## 1.0.7
1101
1102- Renamed /allowed-tools -> /permissions
1103- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1104- Deprecated claude config commands in favor of editing settings.json
1105- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1106- Improved error handling for /install-github-app
1107- Bugfixes, UI polish, and tool reliability improvements
1108
1109## 1.0.6
1110
1111- Improved edit reliability for tab-indented files
1112- Respect CLAUDE_CONFIG_DIR everywhere
1113- Reduced unnecessary tool permission prompts
1114- Added support for symlinks in @file typeahead
1115- Bugfixes, UI polish, and tool reliability improvements
1116
1117## 1.0.4
1118
1119- Fixed a bug where MCP tool errors weren't being parsed correctly
1120
1121## 1.0.1
1122
1123- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1124- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1125- Updated documentation links and OAuth process descriptions
1126
1127## 1.0.0
1128
1129- Claude Code is now generally available
1130- Introducing Sonnet 4 and Opus 4 models
1131
1132## 0.2.125
1133
1134- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1135- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1136
1137## 0.2.117
1138
1139- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1140- Introduced settings.cleanupPeriodDays
1141- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1142- Introduced --debug mode
1143
1144## 0.2.108
1145
1146- You can now send messages to Claude while it works to steer Claude in real-time
1147- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1148- Fixed a bug where thinking was not working in -p mode
1149- Fixed a regression in /cost reporting
1150- Deprecated MCP wizard interface in favor of other MCP commands
1151- Lots of other bugfixes and improvements
1152
1153## 0.2.107
1154
1155- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1156
1157## 0.2.106
1158
1159- MCP SSE server configs can now specify custom headers
1160- Fixed a bug where MCP permission prompt didn't always show correctly
1161
1162## 0.2.105
1163
1164- Claude can now search the web
1165- Moved system & account status to /status
1166- Added word movement keybindings for Vim
1167- Improved latency for startup, todo tool, and file edits
1168
1169## 0.2.102
1170
1171- Improved thinking triggering reliability
1172- Improved @mention reliability for images and folders
1173- You can now paste multiple large chunks into one prompt
1174
1175## 0.2.100
1176
1177- Fixed a crash caused by a stack overflow error
1178- Made db storage optional; missing db support disables --continue and --resume
1179
1180## 0.2.98
1181
1182- Fixed an issue where auto-compact was running twice
1183
1184## 0.2.96
1185
1186- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1187
1188## 0.2.93
1189
1190- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1191- Claude now has access to a Todo list that helps it stay on track and be more organized
1192
1193## 0.2.82
1194
1195- Added support for --disallowedTools
1196- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1197
1198## 0.2.75
1199
1200- Hit Enter to queue up additional messages while Claude is working
1201- Drag in or copy/paste image files directly into the prompt
1202- @-mention files to directly add them to context
1203- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1204- Improved performance for filename auto-complete
1205
1206## 0.2.74
1207
1208- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1209- Task tool can now perform writes and run bash commands
1210
1211## 0.2.72
1212
1213- Updated spinner to indicate tokens loaded and tool usage
1214
1215## 0.2.70
1216
1217- Network commands like curl are now available for Claude to use
1218- Claude can now run multiple web queries in parallel
1219- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1220
1221## 0.2.69
1222
1223- Fixed UI glitches with improved Select component behavior
1224- Enhanced terminal output display with better text truncation logic
1225
1226## 0.2.67
1227
1228- Shared project permission rules can be saved in .claude/settings.json
1229
1230## 0.2.66
1231
1232- Print mode (-p) now supports streaming output via --output-format=stream-json
1233- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1234
1235## 0.2.63
1236
1237- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1238
1239## 0.2.61
1240
1241- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1242- Enhanced image detection for more reliable clipboard paste functionality
1243- Fixed an issue where ESC key could crash the conversation history selector
1244
1245## 0.2.59
1246
1247- Copy+paste images directly into your prompt
1248- Improved progress indicators for bash and fetch tools
1249- Bugfixes for non-interactive mode (-p)
1250
1251## 0.2.54
1252
1253- Quickly add to Memory by starting your message with '#'
1254- Press ctrl+r to see full output for long tool results
1255- Added support for MCP SSE transport
1256
1257## 0.2.53
1258
1259- New web fetch tool lets Claude view URLs that you paste in
1260- Fixed a bug with JPEG detection
1261
1262## 0.2.50
1263
1264- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1265
1266## 0.2.49
1267
1268- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1269
1270## 0.2.47
1271
1272- Press Tab to auto-complete file and folder names
1273- Press Shift + Tab to toggle auto-accept for file edits
1274- Automatic conversation compaction for infinite conversation length (toggle with /config)
1275
1276## 0.2.44
1277
1278- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1279
1280## 0.2.41
1281
1282- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1283- MCP server startup no longer blocks the app from starting up
1284
1285## 0.2.37
1286
1287- New /release-notes command lets you view release notes at any time
1288- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1289
1290## 0.2.36
1291
1292- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1293- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1294
1295## 0.2.34
1296
1297- Vim bindings for text input - enable with /vim or /config
1298
1299## 0.2.32
1300
1301- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1302- Fix for some PersistentShell issues
1303
1304## 0.2.31
1305
1306- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1307- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1308
1309## 0.2.30
1310
1311- Added ANSI color theme for better terminal compatibility
1312- Fixed issue where slash command arguments weren't being sent properly
1313- (Mac-only) API keys are now stored in macOS Keychain
1314
1315## 0.2.26
1316
1317- New /approved-tools command for managing tool permissions
1318- Word-level diff display for improved code readability
1319- Fuzzy matching for slash commands
1320
1321## 0.2.21
1322
1323- Fuzzy matching for /commands