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 5a91286 · on Jan 16, 2026 History
1,407 lines · 69.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.9
4
5- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
6- Added `plansDirectory` setting to customize where plan files are stored
7- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
8- Added session URL attribution to commits and PRs created from web sessions
9- Added support for `PreToolUse` hooks to return `additionalContext` to the model
10- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
11- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
12- Fixed MCP server reconnection hanging when cached connection promise never resolves
13- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
14
15## 2.1.7
16
17- Added customizable keyboard shortcuts via `~/.claude/keybindings.json`. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
18- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
19- Added ability to provide feedback when accepting permission prompts
20- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
21- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
22- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
23- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
24- Fixed context window blocking limit being calculated using the full context window instead of the effective context window (which reserves space for max output tokens)
25- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
26- Fixed terminal title animation jitter by using fixed-width braille characters
27- Fixed plugins with git submodules not being fully initialized when installed
28- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
29- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
30- Enabled MCP tool search auto mode by default for all users. When MCP tool descriptions exceed 10% of the context window, they are automatically deferred and discovered via the MCPSearch tool instead of being loaded upfront. This reduces context usage for users with many MCP tools configured. Users can disable this by adding `MCPSearch` to `disallowedTools` in their settings.
31- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
32- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
33
34## 2.1.6
35
36- Added search functionality to `/config` command for quickly filtering settings
37- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
38- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
39- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
40- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
41- Added an error display when the editor fails during Ctrl+G
42- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
43- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
44- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
45- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
46- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
47- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
48- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
49- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
50- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
51- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
52- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
53- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
54- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
55- Improved @ autocomplete with icons for different suggestion types and single-line formatting
56- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
57- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
58- Changed terminal title to "Claude Code" on startup for better window identification
59- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
60- [VSCode] Fixed usage indicator not updating after manual compact
61
62## 2.1.5
63
64- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
65
66## 2.1.4
67
68- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
69- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
70
71## 2.1.3
72
73- Merged slash commands and skills, simplifying the mental model with no change in behavior
74- Added release channel (`stable` or `latest`) toggle to `/config`
75- Added detection and warnings for unreachable permission rules, with warnings in `/doctor` and after saving rules that include the source of each rule and actionable fix guidance
76- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
77- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
78- Fixed mismatch between background task count in status bar and items shown in tasks dialog
79- Fixed sub-agents using the wrong model during conversation compaction
80- Fixed web search in sub-agents using incorrect model
81- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
82- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
83- Improved slash command suggestion readability by truncating long descriptions to 2 lines
84- Changed tool hook execution timeout from 60 seconds to 10 minutes
85- [VSCode] Added clickable destination selector for permission requests, allowing you to choose where settings are saved (this project, all projects, shared with team, or session only)
86
87## 2.1.2
88
89- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
90- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
91- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
92- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
93- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
94- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
95- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
96- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
97- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
98- Fixed updates incorrectly claiming another installation is in progress
99- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
100- Fixed remote session URL and teleport being broken when using `/tasks` command
101- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
102- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
103- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
104- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
105- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
106- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
107- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
108- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
109- [SDK] Changed minimum zod peer dependency to ^4.0.0
110- [VSCode] Fixed usage display not updating after manual compact
111
112## 2.1.0
113
114- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
115- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
116- Added support for `agent` field in skills to specify agent type for execution
117- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
118- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
119- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
120- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
121- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
122- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
123- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
124- Fixed Esc key with queued prompts to only move them to input without canceling the running task
125- Reduced permission prompts for complex bash commands
126- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
127- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
128- Fixed plugin path resolution for file-based marketplace sources
129- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
130- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
131- Fixed Claude in Chrome support for WSL environments
132- Fixed Windows native installer silently failing when executable creation fails
133- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
134- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
135- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
136- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
137- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
138- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
139- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
140- Added hooks support for skill and slash command frontmatter
141- 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
142- Added `/plan` command shortcut to enable plan mode directly from the prompt
143- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
144- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
145- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
146- Added support for `once: true` config for hooks
147- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
148- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
149- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
150- Added left/right arrow key navigation for cycling through tabs in dialogs
151- Added real-time thinking block display in Ctrl+O transcript mode
152- Added filepath to full output in background bash task details dialog
153- Added Skills as a separate category in the context visualization
154- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
155- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
156- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
157- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
158- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
159- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
160- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
161- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
162- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
163- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
164- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
165- Fixed commands with `$()` command substitution failing with parse errors
166- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
167- 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)
168- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
169- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
170- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
171- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
172- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
173- Fixed subagents sometimes not inheriting the parent's model by default
174- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
175- Fixed duplicate Bash commands appearing in permission request option labels
176- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
177- Fixed background task completion notifications to appear proactively with bullet point
178- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
179- Fixed cursor disappearing after dismissing permission dialogs
180- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
181- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
182- Fixed images being silently dropped when queueing messages while backgrounding a task
183- Fixed large pasted images failing with "Image was too large" error
184- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
185- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
186- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
187- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
188- Fixed spinner token counter to properly accumulate tokens from subagents during execution
189- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
190- Fixed race condition where LSP tool could return "no server available" during startup
191- Fixed feedback submission hanging indefinitely when network requests timeout
192- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
193- Fixed hook success message showing trailing colon when hook has no output
194- Multiple optimizations to improve startup performance
195- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
196- Improved performance when reading Jupyter notebooks with many cells
197- Improved reliability for piped input like `cat refactor.md | claude`
198- Improved reliability for AskQuestion tool
199- Improved sed in-place edit commands to render as file edits with diff preview
200- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
201- Improved compaction reliability
202- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
203- Improved skills to show progress while executing, displaying tool uses as they happen
204- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
205- Improved skill suggestions to prioritize recently and frequently used skills
206- Improved spinner feedback when waiting for the first response token
207- Improved token count display in spinner to include tokens from background agents
208- Improved incremental output for async agents to give the main thread more control and visibility
209- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
210- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
211- Improved macOS screenshot paste reliability with TIFF format support
212- Improved `/stats` output
213- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
214- Changed "Interrupted" message color from red to grey for a less alarming appearance
215- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
216- Removed underline styling from image reference links
217- [SDK] Changed minimum zod peer dependency to ^4.0.0
218- [VSCode] Added currently selected model name to the context menu
219- [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")
220- [VSCode] Fixed paragraph breaks not rendering in markdown content
221- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
222- [Windows] Fixed issue with improper rendering
223
224## 2.0.76
225
226- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
227
228## 2.0.75
229
230- Minor bugfixes
231
232## 2.0.74
233
234- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
235- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
236- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
237- Added syntax highlighting info to theme picker
238- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
239- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
240- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
241- Fixed a potential crash when syntax highlighting isn't initialized correctly
242- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
243- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
244- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
245- [Windows] Fixed issue with improper rendering
246- [VSCode] Added gift tag pictogram for year-end promotion message
247
248## 2.0.73
249
250- Added clickable `[Image #N]` links that open attached images in the default viewer
251- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
252- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
253- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
254- Fixed slow input history cycling and race condition that could overwrite text after message submission
255- Improved `/theme` command to open theme picker directly
256- Improved theme picker UI
257- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
258- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
259
260## 2.0.72
261
262- 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
263- Reduced terminal flickering
264- Added scannable QR code to mobile app tip for quick app downloads
265- Added loading indicator when resuming conversations for better feedback
266- Fixed `/context` command not respecting custom system prompts in non-interactive mode
267- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
268- Improved @ mention file suggestion speed (~3x faster in git repositories)
269- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
270- Improved settings validation errors to be more prominent
271- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
272
273## 2.0.71
274
275- Added /config toggle to enable/disable prompt suggestions
276- Added `/settings` as an alias for the `/config` command
277- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
278- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
279- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
280- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
281- New syntax highlighting engine for native build
282
283## 2.0.70
284
285- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
286- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
287- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
288- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
289- Fixed input being cleared when processing queued commands while the user was typing
290- Fixed prompt suggestions replacing typed input when pressing Tab
291- Fixed diff view not updating when terminal is resized
292- Improved memory usage by 3x for large conversations
293- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
294- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
295- Fix thinking mode toggle in /config not persisting correctly
296- Improve UI for file creation permission dialog
297
298## 2.0.69
299
300- Minor bugfixes
301
302## 2.0.68
303
304- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
305- Fixed a bug where disallowed MCP tools were visible to the model
306- Fixed an issue where steering messages could be lost while a subagent is working
307- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
308- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
309- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
310
311## 2.0.67
312
313- Thinking mode is now enabled by default for Opus 4.5
314- Thinking mode configuration has moved to /config
315- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
316- Show reason why autoupdater is disabled in `/doctor`
317- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
318- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
319- Fixed scroll position resetting after deleting a permission rule in `/permissions`
320- 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
321- Fixed `claude install --force` not bypassing stale lock files
322- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
323- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
324
325## 2.0.65
326
327- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
328- Added context window information to status line input
329- Added `fileSuggestion` setting for custom `@` file search commands
330- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
331- Fixed prompt not being saved to history when aborting a query with Escape
332- Fixed Read tool image handling to identify format from bytes instead of file extension
333
334## 2.0.64
335
336- Made auto-compacting instant
337- Agents and bash commands can run asynchronously and send messages to wake up the main agent
338- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
339- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
340- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
341- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
342- Fixed auto-loading .env when using native installer
343- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
344- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
345- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
346- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
347- Bedrock: Improve efficiency of token counting
348- Bedrock: Add support for `aws login` AWS Management Console credentials
349- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
350
351## 2.0.62
352
353- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
354- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
355- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
356- Fixed slash command selection not working when multiple commands share the same name
357- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
358- Fixed running versions getting removed because lock file incorrectly going stale
359- Fixed IDE diff tab not closing when rejecting file changes
360
361## 2.0.61
362
363- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
364
365## 2.0.60
366
367- Added background agent support. Agents run in the background while you work
368- Added --disable-slash-commands CLI flag to disable all slash commands
369- Added model name to "Co-Authored-By" commit messages
370- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
371- Updated Fetch to skip summarization for pre-approved websites
372- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
373
374## 2.0.59
375
376- Added --agent CLI flag to override the agent setting for the current session
377- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
378- VS Code: Fixed .claude.json config file being read from incorrect location
379
380## 2.0.58
381
382- Pro users now have access to Opus 4.5 as part of their subscription!
383- Fixed timer duration showing "11m 60s" instead of "12m 0s"
384- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
385
386## 2.0.57
387
388- Added feedback input when rejecting plans, allowing users to tell Claude what to change
389- VSCode: Added streaming message support for real-time response display
390
391## 2.0.56
392
393- Added setting to enable/disable terminal progress bar (OSC 9;4)
394- 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.
395
396## 2.0.55
397
398- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
399- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
400- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
401- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
402
403## 2.0.54
404
405- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
406- Fix issue with excessive iTerm notifications
407
408## 2.0.52
409
410- Fixed duplicate message display when starting Claude with a command line argument
411- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
412- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
413- Permit some uses of `$!` in bash commands
414
415## 2.0.51
416
417- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
418- Introducing Claude Code for Desktop: https://claude.com/download
419- 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
420- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
421- Plan Mode now builds more precise plans and executes more thoroughly
422- Usage limit notifications now easier to understand
423- Switched `/usage` back to "% used"
424- Fixed handling of thinking errors
425- Fixed performance regression
426
427## 2.0.50
428
429- Fixed bug preventing calling MCP tools that have nested references in their input schemas
430- Silenced a noisy but harmless error during upgrades
431- Improved ultrathink text display
432- Improved clarity of 5-hour session limit warning message
433
434## 2.0.49
435
436- Added readline-style ctrl-y for pasting deleted text
437- Improved clarity of usage limit warning message
438- Fixed handling of subagent permissions
439
440## 2.0.47
441
442- Improved error messages and validation for `claude --teleport`
443- Improved error handling in `/usage`
444- Fixed race condition with history entry not getting logged at exit
445- Fixed Vertex AI configuration not being applied from `settings.json`
446
447## 2.0.46
448
449- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
450
451## 2.0.45
452
453- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
454- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
455- Send background tasks to Claude Code on the web by starting a message with `&`
456
457## 2.0.43
458
459- Added `permissionMode` field for custom agents
460- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
461- Added skills frontmatter field to declare skills to auto-load for subagents
462- Added the `SubagentStart` hook event
463- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
464- Fixed duplicate rendering of some messages in the UI
465- Fixed some visual flickers
466- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
467
468## 2.0.42
469
470- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
471
472## 2.0.41
473
474- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
475- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
476- Fixed incorrect labeling of user settings vs project settings in command descriptions
477- Fixed crash when plugin command hooks timeout during execution
478- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
479- Fixed broken security documentation links in trust dialogs and onboarding
480- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
481- ctrl-r history search landing on a slash command no longer cancels the search
482- SDK: Support custom timeouts for hooks
483- Allow more safe git commands to run without approval
484- Plugins: Added support for sharing and installing output styles
485- Teleporting a session from web will automatically set the upstream branch
486
487## 2.0.37
488
489- Fixed how idleness is computed for notifications
490- Hooks: Added matcher values for Notification hook events
491- Output Styles: Added `keep-coding-instructions` option to frontmatter
492
493## 2.0.36
494
495- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
496- Fixed queued messages being incorrectly executed as bash commands
497- Fixed input being lost when typing while a queued message is processed
498
499## 2.0.35
500
501- Improve fuzzy search results when searching commands
502- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
503- 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
504- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
505- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
506
507## 2.0.34
508
509- VSCode Extension: Added setting to configure the initial permission mode for new conversations
510- Improved file path suggestion performance with native Rust-based fuzzy finder
511- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
512- Fixed memory crash when reading or writing large files (especially base64-encoded images)
513
514## 2.0.33
515
516- Native binary installs now launch quicker.
517- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
518- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
519
520## 2.0.32
521
522- Un-deprecate output styles based on community feedback
523- Added `companyAnnouncements` setting for displaying announcements on startup
524- Fixed hook progress messages not updating correctly during PostToolUse hook execution
525
526## 2.0.31
527
528- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
529- Vertex: add support for Web Search on supported models
530- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
531- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
532- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
533- Fixed plugin uninstall not removing plugins
534
535## 2.0.30
536
537- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
538- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
539- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
540- Added prompt-based stop hooks
541- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
542- Enabled SSE MCP servers on native build
543- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
544- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
545- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
546- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
547- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
548- Fixed bug that saved session permissions to local settings
549- Fixed MCP tools not being available to sub-agents
550- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
551- Fixed delay when navigating through typeahead suggestions with arrow keys
552- VSCode: Restored selection indicator in input footer showing current file or code selection status
553
554## 2.0.28
555
556- Plan mode: introduced new Plan subagent
557- Subagents: claude can now choose to resume subagents
558- Subagents: claude can dynamically choose the model used by its subagents
559- SDK: added --max-budget-usd flag
560- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
561- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
562- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
563- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
564- Various other bug fixes
565
566## 2.0.27
567
568- New UI for permission prompts
569- Added current branch filtering and search to session resume screen for easier navigation
570- Fixed directory @-mention causing "No assistant message found" error
571- VSCode Extension: Add config setting to include .gitignored files in file searches
572- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
573
574## 2.0.25
575
576- 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
577
578## 2.0.24
579
580- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
581- Claude Code Web: Support for Web -> CLI teleport
582- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
583- Bedrock: Display awsAuthRefresh output when auth is required
584
585## 2.0.22
586
587- Fixed content layout shift when scrolling through slash commands
588- IDE: Add toggle to enable/disable thinking.
589- Fix bug causing duplicate permission prompts with parallel tool calls
590- Add support for enterprise managed MCP allowlist and denylist
591
592## 2.0.21
593
594- Support MCP `structuredContent` field in tool responses
595- Added an interactive question tool
596- Claude will now ask you questions more often in plan mode
597- Added Haiku 4.5 as a model option for Pro users
598- Fixed an issue where queued commands don't have access to previous messages' output
599
600## 2.0.20
601
602- Added support for Claude Skills
603
604## 2.0.19
605
606- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
607- Fixed a bug where Haiku was unnecessarily called in print mode
608
609## 2.0.17
610
611- Added Haiku 4.5 to model selector!
612- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
613- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
614- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
615- OTEL: support HTTP_PROXY and HTTPS_PROXY
616- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
617
618## 2.0.15
619
620- Fixed bug with resuming where previously created files needed to be read again before writing
621- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
622
623## 2.0.14
624
625- Fix @-mentioning MCP servers to toggle them on/off
626- Improve permission checks for bash with inline env vars
627- Fix ultrathink + thinking toggle
628- Reduce unnecessary logins
629- Document --system-prompt
630- Several improvements to rendering
631- Plugins UI polish
632
633## 2.0.13
634
635- Fixed `/plugin` not working on native build
636
637## 2.0.12
638
639- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
640- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
641- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
642- `/plugin validate` command for validating plugin structure and configuration
643- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
644- Plugin documentation available at https://code.claude.com/docs/en/plugins
645- Comprehensive error messages and diagnostics via `/doctor` command
646- Avoid flickering in `/model` selector
647- Improvements to `/help`
648- Avoid mentioning hooks in `/resume` summaries
649- Changes to the "verbose" setting in `/config` now persist across sessions
650
651## 2.0.11
652
653- Reduced system prompt size by 1.4k tokens
654- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
655- Fixed Opus fallback rate limit errors appearing incorrectly
656- Fixed /add-dir command selecting wrong default tab
657
658## 2.0.10
659
660- Rewrote terminal renderer for buttery smooth UI
661- Enable/disable MCP servers by @mentioning, or in /mcp
662- Added tab completion for shell commands in bash mode
663- PreToolUse hooks can now modify tool inputs
664- Press Ctrl-G to edit your prompt in your system's configured text editor
665- Fixes for bash permission checks with environment variables in the command
666
667## 2.0.9
668
669- Fix regression where bash backgrounding stopped working
670
671## 2.0.8
672
673- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
674- IDE: Add drag-and-drop support for files and folders in chat
675- /context: Fix counting for thinking blocks
676- Improve message rendering for users with light themes on dark terminals
677- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
678
679## 2.0.5
680
681- IDE: Fix IME unintended message submission with Enter and Tab
682- IDE: Add "Open in Terminal" link in login screen
683- Fix unhandled OAuth expiration 401 API errors
684- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
685
686## 2.0.1
687
688- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
689- Various bug fixes and presentation improvements
690
691## 2.0.0
692
693- New native VS Code extension
694- Fresh coat of paint throughout the whole app
695- /rewind a conversation to undo code changes
696- /usage command to see plan limits
697- Tab to toggle thinking (sticky across sessions)
698- Ctrl-R to search history
699- Unshipped claude config command
700- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
701- SDK: The Claude Code SDK is now the Claude Agent SDK
702- Add subagents dynamically with `--agents` flag
703
704## 1.0.126
705
706- Enable /context command for Bedrock and Vertex
707- Add mTLS support for HTTP-based OpenTelemetry exporters
708
709## 1.0.124
710
711- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
712- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
713- No longer inform Claude of the list of allowed tools when permission is denied
714- Fixed security vulnerability in Bash tool permission checks
715- Improved VSCode extension performance for large files
716
717## 1.0.123
718
719- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
720- Fixed thinking mode triggering on negation phrases like "don't think"
721- Fixed rendering performance degradation during token streaming
722- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
723- Enhanced BashTool environment snapshot logging
724- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
725- Migrated --debug logging to a file, to enable easy tailing & filtering
726
727## 1.0.120
728
729- Fix input lag during typing, especially noticeable with large prompts
730- Improved VSCode extension command registry and sessions dialog user experience
731- Enhanced sessions dialog responsiveness and visual feedback
732- Fixed IDE compatibility issue by removing worktree support check
733- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
734
735## 1.0.119
736
737- Fix Windows issue where process visually freezes on entering interactive mode
738- Support dynamic headers for MCP servers via headersHelper configuration
739- Fix thinking mode not working in headless sessions
740- Fix slash commands now properly update allowed tools instead of replacing them
741
742## 1.0.117
743
744- Add Ctrl-R history search to recall previous commands like bash/zsh
745- Fix input lag while typing, especially on Windows
746- Add sed command to auto-allowed commands in acceptEdits mode
747- Fix Windows PATH comparison to be case-insensitive for drive letters
748- Add permissions management hint to /add-dir output
749
750## 1.0.115
751
752- Improve thinking mode display with enhanced visual effects
753- Type /t to temporarily disable thinking mode in your prompt
754- Improve path validation for glob and grep tools
755- Show condensed output for post-tool hooks to reduce visual clutter
756- Fix visual feedback when loading state completes
757- Improve UI consistency for permission request dialogs
758
759## 1.0.113
760
761- Deprecated piped input in interactive mode
762- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
763
764## 1.0.112
765
766- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
767- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
768- Hooks: Added systemMessage support for SessionEnd hooks
769- Added `spinnerTipsEnabled` setting to disable spinner tips
770- IDE: Various improvements and bug fixes
771
772## 1.0.111
773
774- /model now validates provided model names
775- Fixed Bash tool crashes caused by malformed shell syntax parsing
776
777## 1.0.110
778
779- /terminal-setup command now supports WezTerm
780- MCP: OAuth tokens now proactively refresh before expiration
781- Fixed reliability issues with background Bash processes
782
783## 1.0.109
784
785- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
786
787## 1.0.106
788
789- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
790
791## 1.0.97
792
793- Settings: /doctor now validates permission rule syntax and suggests corrections
794
795## 1.0.94
796
797- Vertex: add support for global endpoints for supported models
798- /memory command now allows direct editing of all imported memory files
799- SDK: Add custom tools as callbacks
800- Added /todos command to list current todo items
801
802## 1.0.93
803
804- Windows: Add alt + v shortcut for pasting images from clipboard
805- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
806
807## 1.0.90
808
809- Settings file changes take effect immediately - no restart required
810
811## 1.0.88
812
813- Fixed issue causing "OAuth authentication is currently not supported"
814- Status line input now includes `exceeds_200k_tokens`
815- Fixed incorrect usage tracking in /cost.
816- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
817- Bedrock: Updated default Sonnet model to Sonnet 4
818
819## 1.0.86
820
821- Added /context to help users self-serve debug context issues
822- SDK: Added UUID support for all SDK messages
823- SDK: Added `--replay-user-messages` to replay user messages back to stdout
824
825## 1.0.85
826
827- Status line input now includes session cost info
828- Hooks: Introduced SessionEnd hook
829
830## 1.0.84
831
832- Fix tool_use/tool_result id mismatch error when network is unstable
833- Fix Claude sometimes ignoring real-time steering when wrapping up a task
834- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
835- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
836
837## 1.0.83
838
839- @-mention: Support files with spaces in path
840- New shimmering spinner
841
842## 1.0.82
843
844- SDK: Add request cancellation support
845- SDK: New additionalDirectories option to search custom paths, improved slash command processing
846- Settings: Validation prevents invalid fields in .claude/settings.json files
847- MCP: Improve tool name consistency
848- Bash: Fix crash when Claude tries to automatically read large files
849
850## 1.0.81
851
852- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
853- Agents: Fix custom agent loading when agent files are unparsable
854
855## 1.0.80
856
857- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
858
859## 1.0.77
860
861- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
862- SDK: Add session support and permission denial tracking
863- Fix token limit errors in conversation summarization
864- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
865
866## 1.0.73
867
868- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
869- MCP: Press Esc to cancel OAuth authentication flows
870- Bash: Improved command validation and reduced false security warnings
871- UI: Enhanced spinner animations and status line visual hierarchy
872- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
873
874## 1.0.72
875
876- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
877
878## 1.0.71
879
880- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
881- Customizable status line: add your terminal prompt to Claude Code with /statusline
882
883## 1.0.70
884
885- Performance: Optimized message rendering for better performance with large contexts
886- Windows: Fixed native file search, ripgrep, and subagent functionality
887- Added support for @-mentions in slash command arguments
888
889## 1.0.69
890
891- Upgraded Opus to version 4.1
892
893## 1.0.68
894
895- Fix incorrect model names being used for certain commands like `/pr-comments`
896- 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.
897- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
898- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
899- SDK: Added canUseTool callback support for tool confirmation
900- Added `disableAllHooks` setting
901- Improved file suggestions performance in large repos
902
903## 1.0.65
904
905- IDE: Fixed connection stability issues and error handling for diagnostics
906- Windows: Fixed shell environment setup for users without .bashrc files
907
908## 1.0.64
909
910- Agents: Added model customization support - you can now specify which model an agent should use
911- Agents: Fixed unintended access to the recursive agent tool
912- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
913- SDK: Fixed user input tracking across multi-turn conversations
914- Added hidden files to file search and @-mention suggestions
915
916## 1.0.63
917
918- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
919
920## 1.0.62
921
922- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
923- Hooks: Added SessionStart hook for new session initialization
924- /add-dir command now supports typeahead for directory paths
925- Improved network connectivity check reliability
926
927## 1.0.61
928
929- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
930- Settings: Added `--settings` flag to load settings from a JSON file
931- Settings: Fixed resolution of settings files paths that are symlinks
932- OTEL: Fixed reporting of wrong organization after authentication changes
933- Slash commands: Fixed permissions checking for allowed-tools with Bash
934- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
935- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
936- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
937
938## 1.0.60
939
940- You can now create custom subagents for specialized tasks! Run /agents to get started
941
942## 1.0.59
943
944- SDK: Added tool confirmation support with canUseTool callback
945- SDK: Allow specifying env for spawned process
946- Hooks: Exposed PermissionDecision to hooks (including "ask")
947- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
948- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
949
950## 1.0.58
951
952- Added support for reading PDFs
953- MCP: Improved server health status display in 'claude mcp list'
954- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
955
956## 1.0.57
957
958- Added support for specifying a model in slash commands
959- Improved permission messages to help Claude understand allowed tools
960- Fix: Remove trailing newlines from bash output in terminal wrapping
961
962## 1.0.56
963
964- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
965- Fixes for WSL IDE detection
966- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
967
968## 1.0.55
969
970- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
971- Windows: fixed Ctrl+Z crash
972- SDK: Added ability to capture error logging
973- Add --system-prompt-file option to override system prompt in print mode
974
975## 1.0.54
976
977- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
978- Custom slash commands: Added argument-hint to frontmatter
979- Windows: OAuth uses port 45454 and properly constructs browser URL
980- Windows: mode switching now uses alt + m, and plan mode renders properly
981- Shell: Switch to in-memory shell snapshot to fix file-related errors
982
983## 1.0.53
984
985- Updated @-mention file truncation from 100 lines to 2000 lines
986- 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).
987
988## 1.0.52
989
990- Added support for MCP server instructions
991
992## 1.0.51
993
994- Added support for native Windows (requires Git for Windows)
995- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
996- Settings: /doctor can now help you identify and fix invalid setting files
997- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
998- Increased auto-compact warning threshold from 60% to 80%
999- Fixed an issue with handling user directories with spaces for shell snapshots
1000- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1001- Custom slash commands: Fixed user-level commands in subdirectories
1002- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1003
1004## 1.0.48
1005
1006- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1007- Added progress messages to Bash tool based on the last 5 lines of command output
1008- Added expanding variables support for MCP server configuration
1009- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1010- Improved IDE extension path handling when Claude Code runs in WSL
1011- Hooks: Added a PreCompact hook
1012- Vim mode: Added c, f/F, t/T
1013
1014## 1.0.45
1015
1016- Redesigned Search (Grep) tool with new tool input parameters and features
1017- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1018- Fixed config file corruption issue by enforcing atomic writes
1019- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1020- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1021- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1022
1023## 1.0.44
1024
1025- New /export command lets you quickly export a conversation for sharing
1026- MCP: resource_link tool results are now supported
1027- MCP: tool annotations and tool titles now display in /mcp view
1028- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1029
1030## 1.0.43
1031
1032- Fixed a bug where the theme selector was saving excessively
1033- Hooks: Added EPIPE system error handling
1034
1035## 1.0.42
1036
1037- Added tilde (`~`) expansion support to `/add-dir` command
1038
1039## 1.0.41
1040
1041- Hooks: Split Stop hook triggering into Stop and SubagentStop
1042- Hooks: Enabled optional timeout configuration for each command
1043- Hooks: Added "hook_event_name" to hook input
1044- Fixed a bug where MCP tools would display twice in tool list
1045- New tool parameters JSON for Bash tool in `tool_decision` event
1046
1047## 1.0.40
1048
1049- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1050
1051## 1.0.39
1052
1053- New Active Time metric in OpenTelemetry logging
1054
1055## 1.0.38
1056
1057- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1058
1059## 1.0.37
1060
1061- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1062
1063## 1.0.36
1064
1065- Web search now takes today's date into context
1066- Fixed a bug where stdio MCP servers were not terminating properly on exit
1067
1068## 1.0.35
1069
1070- Added support for MCP OAuth Authorization Server discovery
1071
1072## 1.0.34
1073
1074- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1075
1076## 1.0.33
1077
1078- Improved logging functionality with session ID support
1079- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1080- Improvements to plan mode
1081
1082## 1.0.32
1083
1084- Updated loopback config for litellm
1085- Added forceLoginMethod setting to bypass login selection screen
1086
1087## 1.0.31
1088
1089- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1090
1091## 1.0.30
1092
1093- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1094- Improved file path autocomplete with filename matching
1095- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1096- Enhanced jq regex support for complex filters with pipes and select
1097
1098## 1.0.29
1099
1100- Improved CJK character support in cursor navigation and rendering
1101
1102## 1.0.28
1103
1104- Slash commands: Fix selector display during history navigation
1105- Resizes images before upload to prevent API size limit errors
1106- Added XDG_CONFIG_HOME support to configuration directory
1107- Performance optimizations for memory usage
1108- New attributes (terminal.type, language) in OpenTelemetry logging
1109
1110## 1.0.27
1111
1112- Streamable HTTP MCP servers are now supported
1113- Remote MCP servers (SSE and HTTP) now support OAuth
1114- MCP resources can now be @-mentioned
1115- /resume slash command to switch conversations within Claude Code
1116
1117## 1.0.25
1118
1119- Slash commands: moved "project" and "user" prefixes to descriptions
1120- Slash commands: improved reliability for command discovery
1121- Improved support for Ghostty
1122- Improved web search reliability
1123
1124## 1.0.24
1125
1126- Improved /mcp output
1127- Fixed a bug where settings arrays got overwritten instead of merged
1128
1129## 1.0.23
1130
1131- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1132- Released Python SDK: pip install claude-code-sdk to get started
1133
1134## 1.0.22
1135
1136- SDK: Renamed `total_cost` to `total_cost_usd`
1137
1138## 1.0.21
1139
1140- Improved editing of files with tab-based indentation
1141- Fix for tool_use without matching tool_result errors
1142- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1143
1144## 1.0.18
1145
1146- Added --add-dir CLI argument for specifying additional working directories
1147- Added streaming input support without require -p flag
1148- Improved startup performance and session storage performance
1149- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1150- Added detailed MCP server tools display (/mcp)
1151- MCP authentication and permission improvements
1152- Added auto-reconnection for MCP SSE connections on disconnect
1153- Fixed issue where pasted content was lost when dialogs appeared
1154
1155## 1.0.17
1156
1157- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1158- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1159- MCP server list UI improvements
1160- Update Claude Code process title to display "claude" instead of "node"
1161
1162## 1.0.11
1163
1164- Claude Code can now also be used with a Claude Pro subscription
1165- Added /upgrade for smoother switching to Claude Max plans
1166- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1167- Improved shell configuration error handling
1168- Improved todo list handling during compaction
1169
1170## 1.0.10
1171
1172- Added markdown table support
1173- Improved streaming performance
1174
1175## 1.0.8
1176
1177- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1178- Increased default otel interval from 1s -> 5s
1179- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1180- Fixed a regression where search tools unnecessarily asked for permissions
1181- Added support for triggering thinking non-English languages
1182- Improved compacting UI
1183
1184## 1.0.7
1185
1186- Renamed /allowed-tools -> /permissions
1187- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1188- Deprecated claude config commands in favor of editing settings.json
1189- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1190- Improved error handling for /install-github-app
1191- Bugfixes, UI polish, and tool reliability improvements
1192
1193## 1.0.6
1194
1195- Improved edit reliability for tab-indented files
1196- Respect CLAUDE_CONFIG_DIR everywhere
1197- Reduced unnecessary tool permission prompts
1198- Added support for symlinks in @file typeahead
1199- Bugfixes, UI polish, and tool reliability improvements
1200
1201## 1.0.4
1202
1203- Fixed a bug where MCP tool errors weren't being parsed correctly
1204
1205## 1.0.1
1206
1207- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1208- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1209- Updated documentation links and OAuth process descriptions
1210
1211## 1.0.0
1212
1213- Claude Code is now generally available
1214- Introducing Sonnet 4 and Opus 4 models
1215
1216## 0.2.125
1217
1218- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1219- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1220
1221## 0.2.117
1222
1223- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1224- Introduced settings.cleanupPeriodDays
1225- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1226- Introduced --debug mode
1227
1228## 0.2.108
1229
1230- You can now send messages to Claude while it works to steer Claude in real-time
1231- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1232- Fixed a bug where thinking was not working in -p mode
1233- Fixed a regression in /cost reporting
1234- Deprecated MCP wizard interface in favor of other MCP commands
1235- Lots of other bugfixes and improvements
1236
1237## 0.2.107
1238
1239- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1240
1241## 0.2.106
1242
1243- MCP SSE server configs can now specify custom headers
1244- Fixed a bug where MCP permission prompt didn't always show correctly
1245
1246## 0.2.105
1247
1248- Claude can now search the web
1249- Moved system & account status to /status
1250- Added word movement keybindings for Vim
1251- Improved latency for startup, todo tool, and file edits
1252
1253## 0.2.102
1254
1255- Improved thinking triggering reliability
1256- Improved @mention reliability for images and folders
1257- You can now paste multiple large chunks into one prompt
1258
1259## 0.2.100
1260
1261- Fixed a crash caused by a stack overflow error
1262- Made db storage optional; missing db support disables --continue and --resume
1263
1264## 0.2.98
1265
1266- Fixed an issue where auto-compact was running twice
1267
1268## 0.2.96
1269
1270- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1271
1272## 0.2.93
1273
1274- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1275- Claude now has access to a Todo list that helps it stay on track and be more organized
1276
1277## 0.2.82
1278
1279- Added support for --disallowedTools
1280- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1281
1282## 0.2.75
1283
1284- Hit Enter to queue up additional messages while Claude is working
1285- Drag in or copy/paste image files directly into the prompt
1286- @-mention files to directly add them to context
1287- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1288- Improved performance for filename auto-complete
1289
1290## 0.2.74
1291
1292- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1293- Task tool can now perform writes and run bash commands
1294
1295## 0.2.72
1296
1297- Updated spinner to indicate tokens loaded and tool usage
1298
1299## 0.2.70
1300
1301- Network commands like curl are now available for Claude to use
1302- Claude can now run multiple web queries in parallel
1303- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1304
1305## 0.2.69
1306
1307- Fixed UI glitches with improved Select component behavior
1308- Enhanced terminal output display with better text truncation logic
1309
1310## 0.2.67
1311
1312- Shared project permission rules can be saved in .claude/settings.json
1313
1314## 0.2.66
1315
1316- Print mode (-p) now supports streaming output via --output-format=stream-json
1317- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1318
1319## 0.2.63
1320
1321- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1322
1323## 0.2.61
1324
1325- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1326- Enhanced image detection for more reliable clipboard paste functionality
1327- Fixed an issue where ESC key could crash the conversation history selector
1328
1329## 0.2.59
1330
1331- Copy+paste images directly into your prompt
1332- Improved progress indicators for bash and fetch tools
1333- Bugfixes for non-interactive mode (-p)
1334
1335## 0.2.54
1336
1337- Quickly add to Memory by starting your message with '#'
1338- Press ctrl+r to see full output for long tool results
1339- Added support for MCP SSE transport
1340
1341## 0.2.53
1342
1343- New web fetch tool lets Claude view URLs that you paste in
1344- Fixed a bug with JPEG detection
1345
1346## 0.2.50
1347
1348- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1349
1350## 0.2.49
1351
1352- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1353
1354## 0.2.47
1355
1356- Press Tab to auto-complete file and folder names
1357- Press Shift + Tab to toggle auto-accept for file edits
1358- Automatic conversation compaction for infinite conversation length (toggle with /config)
1359
1360## 0.2.44
1361
1362- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1363
1364## 0.2.41
1365
1366- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1367- MCP server startup no longer blocks the app from starting up
1368
1369## 0.2.37
1370
1371- New /release-notes command lets you view release notes at any time
1372- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1373
1374## 0.2.36
1375
1376- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1377- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1378
1379## 0.2.34
1380
1381- Vim bindings for text input - enable with /vim or /config
1382
1383## 0.2.32
1384
1385- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1386- Fix for some PersistentShell issues
1387
1388## 0.2.31
1389
1390- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1391- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1392
1393## 0.2.30
1394
1395- Added ANSI color theme for better terminal compatibility
1396- Fixed issue where slash command arguments weren't being sent properly
1397- (Mac-only) API keys are now stored in macOS Keychain
1398
1399## 0.2.26
1400
1401- New /approved-tools command for managing tool permissions
1402- Word-level diff display for improved code readability
1403- Fuzzy matching for slash commands
1404
1405## 0.2.21
1406
1407- Fuzzy matching for /commands