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 32c7ff2 · on Feb 17, 2026 History
1,712 lines · 90.0 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.44
4
5- Fixed ENAMETOOLONG errors for deeply-nested directory paths
6- Fixed auth refresh errors
7
8## 2.1.43
9
10- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
11- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
12- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
13
14## 2.1.42
15
16- Improved startup performance by deferring Zod schema construction
17- Improved prompt cache hit rates by moving date out of system prompt
18- Added one-time Opus 4.6 effort callout for eligible users
19- Fixed /resume showing interrupt messages as session titles
20- Fixed image dimension limit errors to suggest /compact
21
22## 2.1.41
23
24- Added guard against launching Claude Code inside another Claude Code session
25- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
26- Fixed a crash when MCP tools return image content during streaming
27- Fixed /resume session previews showing raw XML tags instead of readable command names
28- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
29- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
30- Fixed hook blocking errors (exit code 2) not showing stderr to the user
31- Added `speed` attribute to OTel events and trace spans for fast mode visibility
32- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
33- Added Windows ARM64 (win32-arm64) native binary support
34- Improved `/rename` to auto-generate session name from conversation context when called without arguments
35- Improved narrow terminal layout for prompt footer
36- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
37- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
38- Fixed background task notifications not being delivered in streaming Agent SDK mode
39- Fixed cursor jumping to end on each keystroke in classifier rule input
40- Fixed markdown link display text being dropped for raw URL
41- Fixed auto-compact failure error notifications being shown to users
42- Fixed permission wait time being included in subagent elapsed time display
43- Fixed proactive ticks firing while in plan mode
44- Fixed clear stale permission rules when settings change on disk
45- Fixed hook blocking errors showing stderr content in UI
46
47## 2.1.39
48
49- Improved terminal rendering performance
50- Fixed fatal errors being swallowed instead of displayed
51- Fixed process hanging after session close
52- Fixed character loss at terminal screen boundary
53- Fixed blank lines in verbose transcript view
54
55## 2.1.38
56
57- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
58- Fixed Tab key queueing slash commands instead of autocompleting
59- Fixed bash permission matching for commands using environment variable wrappers
60- Fixed text between tool uses disappearing when not using streaming
61- Fixed duplicate sessions when resuming in VS Code extension
62- Improved heredoc delimiter parsing to prevent command smuggling
63- Blocked writes to `.claude/skills` directory in sandbox mode
64
65## 2.1.37
66
67- Fixed an issue where /fast was not immediately available after enabling /extra-usage
68
69## 2.1.36
70
71- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
72
73## 2.1.34
74
75- Fixed a crash when agent teams setting changed between renders
76- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
77
78## 2.1.33
79
80- Fixed agent teammate sessions in tmux to send and receive messages
81- Fixed warnings about agent teams not being available on your current plan
82- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
83- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
84- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
85- Added plugin name to skill descriptions and `/skills` menu for better discoverability
86- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
87- Fixed an API error that could occur when aborting mid-stream, where whitespace text combined with a thinking block would bypass normalization and produce an invalid request
88- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
89- Fixed an issue where proxy settings configured via `settings.json` environment variables were not applied to WebFetch and other HTTP requests on the Node.js build
90- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
91- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
92- Errors from invalid managed settings are now surfaced
93- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
94- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
95- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
96
97## 2.1.32
98
99- Claude Opus 4.6 is now available!
100- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
101- Claude now automatically records and recalls memories as it works
102- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
103- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
104- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
105- Updated --resume to re-use --agent value specified in previous conversation by default.
106- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
107- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
108- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
109- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
110- VSCode: Added spinner when loading past conversations list
111
112## 2.1.31
113
114- Added session resume hint on exit, showing how to continue your conversation later
115- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
116- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
117- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
118- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
119- Fixed `temperatureOverride` being silently ignored in the streaming API path, causing all streaming requests to use the default temperature (1) regardless of the configured override
120- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
121- Improved system prompts to more clearly guide the model toward using dedicated tools (Read, Edit, Glob, Grep) instead of bash equivalents (`cat`, `sed`, `grep`, `find`), reducing unnecessary bash command usage
122- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
123- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
124- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
125
126## 2.1.30
127
128- Added `pages` parameter to the Read tool for PDFs, allowing specific page ranges to be read (e.g., `pages: "1-5"`). Large PDFs (>10 pages) now return a lightweight reference when `@` mentioned instead of being inlined into context.
129- Added pre-configured OAuth client credentials for MCP servers that don't support Dynamic Client Registration (e.g., Slack). Use `--client-id` and `--client-secret` with `claude mcp add`.
130- Added `/debug` for Claude to help troubleshoot the current session
131- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
132- Added token count, tool uses, and duration metrics to Task tool results
133- Added reduced motion mode to the config
134- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
135- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
136- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
137- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
138- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
139- Fixed permission dialogs stealing focus while actively typing
140- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
141- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
142- Improved memory usage for `--resume` (68% reduction for users with many sessions) by replacing the session index with lightweight stat-based loading and progressive enrichment
143- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
144- Changed `/model` to execute immediately instead of being queued
145- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
146- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
147
148## 2.1.29
149
150- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
151
152## 2.1.27
153
154- Added tool call failures and denials to debug logs
155- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
156- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
157- Sessions are now automatically linked to PRs when created via `gh pr create`
158- Fixed /context command not displaying colored output
159- Fixed status bar duplicating background task indicator when PR status was shown
160- Windows: Fixed bash command execution failing for users with `.bashrc` files
161- Windows: Fixed console windows flashing when spawning child processes
162- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
163
164## 2.1.25
165
166- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
167
168## 2.1.23
169
170- Added customizable spinner verbs setting (`spinnerVerbs`)
171- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
172- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
173- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
174- Fixed pending async hooks not being cancelled when headless streaming sessions ended
175- Fixed tab completion not updating the input field when accepting a suggestion
176- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
177- Improved terminal rendering performance with optimized screen data layout
178- Changed Bash commands to show timeout duration alongside elapsed time
179- Changed merged pull requests to show a purple status indicator in the prompt footer
180- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
181
182## 2.1.22
183
184- Fixed structured outputs for non-interactive (-p) mode
185
186## 2.1.21
187
188- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
189- Fixed shell completion cache files being truncated on exit
190- Fixed API errors when resuming sessions that were interrupted during tool execution
191- Fixed auto-compact triggering too early on models with large output token limits
192- Fixed task IDs potentially being reused after deletion
193- Fixed file search not working in VS Code extension on Windows
194- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
195- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
196- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
197- [VSCode] Fixed message action buttons having incorrect background colors
198
199## 2.1.20
200
201- Added arrow key history navigation in vim normal mode when cursor cannot move further
202- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
203- Added PR review status indicator to the prompt footer, showing the current branch's PR state (approved, changes requested, pending, or draft) as a colored dot with a clickable link
204- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
205- Added ability to delete tasks via the `TaskUpdate` tool
206- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
207- Fixed agents sometimes ignoring user messages sent while actively working on a task
208- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
209- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
210- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
211- Fixed draft prompt being lost when pressing UP arrow to navigate command history
212- Fixed ghost text flickering when typing slash commands mid-input
213- Fixed marketplace source removal not properly deleting settings
214- Fixed duplicate output in some commands like `/context`
215- Fixed task list sometimes showing outside the main conversation view
216- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
217- Fixed crashes when cancelling tool use
218- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
219- Improved thinking status text with a subtle shimmer animation
220- Improved task list to dynamically adjust visible items based on terminal height
221- Improved fork conversation hint to show how to resume the original session
222- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
223- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
224- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
225- Changed the `/copy` command to be available to all users
226- Changed background agents to prompt for tool permissions before launching
227- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
228- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
229
230## 2.1.19
231
232- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
233- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
234- Fixed crashes on processors without AVX instruction support
235- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
236- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
237- Fixed resuming sessions by custom title not working when run from a different directory
238- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
239- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
240- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
241- Fixed file write preview omitting empty lines
242- Changed skills without additional permissions or hooks to be allowed without requiring approval
243- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
244- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
245- [VSCode] Enabled session forking and rewind functionality for all users
246
247## 2.1.18
248
249- Added customizable keyboard shortcuts. Configure keybindings per context, create chord sequences, and personalize your workflow. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
250
251## 2.1.17
252
253- Fixed crashes on processors without AVX instruction support
254
255## 2.1.16
256
257- Added new task management system, including new capabilities like dependency tracking
258- [VSCode] Added native plugin management support
259- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
260- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
261- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
262- Fixed session titles on the resume screen not respecting the user's language setting
263- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
264
265## 2.1.15
266
267- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
268- Improved UI rendering performance with React Compiler
269- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
270- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
271
272## 2.1.14
273
274- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
275- Added search to installed plugins list - type to filter by name or description
276- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
277- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
278- Fixed memory issues that could cause crashes when running parallel subagents
279- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
280- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
281- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
282- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
283- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
284- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
285- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
286- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
287- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
288- Improved backspace to delete pasted text as a single token instead of one character at a time
289- [VSCode] Added `/usage` command to display current plan usage
290
291## 2.1.12
292
293- Fixed message rendering bug
294
295## 2.1.11
296
297- Fixed excessive MCP connection requests for HTTP/SSE transports
298
299## 2.1.10
300
301- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
302- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
303- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
304- Improved startup to capture keystrokes typed before the REPL is fully ready
305- Improved file suggestions to show as removable attachments instead of inserting text when accepted
306- [VSCode] Added install count display to plugin listings
307- [VSCode] Added trust warning when installing plugins
308
309## 2.1.9
310
311- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
312- Added `plansDirectory` setting to customize where plan files are stored
313- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
314- Added session URL attribution to commits and PRs created from web sessions
315- Added support for `PreToolUse` hooks to return `additionalContext` to the model
316- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
317- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
318- Fixed MCP server reconnection hanging when cached connection promise never resolves
319- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
320
321## 2.1.7
322
323- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
324- Added ability to provide feedback when accepting permission prompts
325- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
326- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
327- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
328- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
329- 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)
330- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
331- Fixed terminal title animation jitter by using fixed-width braille characters
332- Fixed plugins with git submodules not being fully initialized when installed
333- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
334- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
335- 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.
336- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
337- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
338
339## 2.1.6
340
341- Added search functionality to `/config` command for quickly filtering settings
342- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
343- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
344- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
345- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
346- Added an error display when the editor fails during Ctrl+G
347- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
348- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
349- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
350- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
351- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
352- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
353- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
354- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
355- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
356- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
357- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
358- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
359- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
360- Improved @ autocomplete with icons for different suggestion types and single-line formatting
361- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
362- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
363- Changed terminal title to "Claude Code" on startup for better window identification
364- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
365- [VSCode] Fixed usage indicator not updating after manual compact
366
367## 2.1.5
368
369- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
370
371## 2.1.4
372
373- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
374- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
375
376## 2.1.3
377
378- Merged slash commands and skills, simplifying the mental model with no change in behavior
379- Added release channel (`stable` or `latest`) toggle to `/config`
380- 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
381- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
382- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
383- Fixed mismatch between background task count in status bar and items shown in tasks dialog
384- Fixed sub-agents using the wrong model during conversation compaction
385- Fixed web search in sub-agents using incorrect model
386- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
387- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
388- Improved slash command suggestion readability by truncating long descriptions to 2 lines
389- Changed tool hook execution timeout from 60 seconds to 10 minutes
390- [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)
391
392## 2.1.2
393
394- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
395- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
396- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
397- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
398- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
399- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
400- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
401- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
402- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
403- Fixed updates incorrectly claiming another installation is in progress
404- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
405- Fixed remote session URL and teleport being broken when using `/tasks` command
406- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
407- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
408- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
409- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
410- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
411- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
412- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
413- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
414- [SDK] Changed minimum zod peer dependency to ^4.0.0
415- [VSCode] Fixed usage display not updating after manual compact
416
417## 2.1.0
418
419- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
420- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
421- Added support for `agent` field in skills to specify agent type for execution
422- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
423- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
424- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
425- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
426- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
427- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
428- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
429- Fixed Esc key with queued prompts to only move them to input without canceling the running task
430- Reduced permission prompts for complex bash commands
431- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
432- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
433- Fixed plugin path resolution for file-based marketplace sources
434- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
435- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
436- Fixed Claude in Chrome support for WSL environments
437- Fixed Windows native installer silently failing when executable creation fails
438- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
439- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
440- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
441- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
442- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
443- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
444- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
445- Added hooks support for skill and slash command frontmatter
446- 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
447- Added `/plan` command shortcut to enable plan mode directly from the prompt
448- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
449- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
450- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
451- Added support for `once: true` config for hooks
452- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
453- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
454- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
455- Added left/right arrow key navigation for cycling through tabs in dialogs
456- Added real-time thinking block display in Ctrl+O transcript mode
457- Added filepath to full output in background bash task details dialog
458- Added Skills as a separate category in the context visualization
459- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
460- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
461- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
462- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
463- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
464- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
465- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
466- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
467- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
468- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
469- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
470- Fixed commands with `$()` command substitution failing with parse errors
471- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
472- 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)
473- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
474- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
475- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
476- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
477- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
478- Fixed subagents sometimes not inheriting the parent's model by default
479- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
480- Fixed duplicate Bash commands appearing in permission request option labels
481- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
482- Fixed background task completion notifications to appear proactively with bullet point
483- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
484- Fixed cursor disappearing after dismissing permission dialogs
485- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
486- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
487- Fixed images being silently dropped when queueing messages while backgrounding a task
488- Fixed large pasted images failing with "Image was too large" error
489- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
490- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
491- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
492- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
493- Fixed spinner token counter to properly accumulate tokens from subagents during execution
494- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
495- Fixed race condition where LSP tool could return "no server available" during startup
496- Fixed feedback submission hanging indefinitely when network requests timeout
497- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
498- Fixed hook success message showing trailing colon when hook has no output
499- Multiple optimizations to improve startup performance
500- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
501- Improved performance when reading Jupyter notebooks with many cells
502- Improved reliability for piped input like `cat refactor.md | claude`
503- Improved reliability for AskQuestion tool
504- Improved sed in-place edit commands to render as file edits with diff preview
505- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
506- Improved compaction reliability
507- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
508- Improved skills to show progress while executing, displaying tool uses as they happen
509- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
510- Improved skill suggestions to prioritize recently and frequently used skills
511- Improved spinner feedback when waiting for the first response token
512- Improved token count display in spinner to include tokens from background agents
513- Improved incremental output for async agents to give the main thread more control and visibility
514- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
515- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
516- Improved macOS screenshot paste reliability with TIFF format support
517- Improved `/stats` output
518- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
519- Changed "Interrupted" message color from red to grey for a less alarming appearance
520- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
521- Removed underline styling from image reference links
522- [SDK] Changed minimum zod peer dependency to ^4.0.0
523- [VSCode] Added currently selected model name to the context menu
524- [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")
525- [VSCode] Fixed paragraph breaks not rendering in markdown content
526- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
527- [Windows] Fixed issue with improper rendering
528
529## 2.0.76
530
531- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
532
533## 2.0.75
534
535- Minor bugfixes
536
537## 2.0.74
538
539- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
540- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
541- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
542- Added syntax highlighting info to theme picker
543- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
544- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
545- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
546- Fixed a potential crash when syntax highlighting isn't initialized correctly
547- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
548- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
549- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
550- [Windows] Fixed issue with improper rendering
551- [VSCode] Added gift tag pictogram for year-end promotion message
552
553## 2.0.73
554
555- Added clickable `[Image #N]` links that open attached images in the default viewer
556- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
557- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
558- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
559- Fixed slow input history cycling and race condition that could overwrite text after message submission
560- Improved `/theme` command to open theme picker directly
561- Improved theme picker UI
562- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
563- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
564
565## 2.0.72
566
567- 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
568- Reduced terminal flickering
569- Added scannable QR code to mobile app tip for quick app downloads
570- Added loading indicator when resuming conversations for better feedback
571- Fixed `/context` command not respecting custom system prompts in non-interactive mode
572- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
573- Improved @ mention file suggestion speed (~3× faster in git repositories)
574- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
575- Improved settings validation errors to be more prominent
576- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
577
578## 2.0.71
579
580- Added /config toggle to enable/disable prompt suggestions
581- Added `/settings` as an alias for the `/config` command
582- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
583- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
584- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
585- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
586- New syntax highlighting engine for native build
587
588## 2.0.70
589
590- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
591- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
592- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
593- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
594- Fixed input being cleared when processing queued commands while the user was typing
595- Fixed prompt suggestions replacing typed input when pressing Tab
596- Fixed diff view not updating when terminal is resized
597- Improved memory usage by 3x for large conversations
598- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
599- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
600- Fix thinking mode toggle in /config not persisting correctly
601- Improve UI for file creation permission dialog
602
603## 2.0.69
604
605- Minor bugfixes
606
607## 2.0.68
608
609- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
610- Fixed a bug where disallowed MCP tools were visible to the model
611- Fixed an issue where steering messages could be lost while a subagent is working
612- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
613- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
614- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
615
616## 2.0.67
617
618- Thinking mode is now enabled by default for Opus 4.5
619- Thinking mode configuration has moved to /config
620- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
621- Show reason why autoupdater is disabled in `/doctor`
622- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
623- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
624- Fixed scroll position resetting after deleting a permission rule in `/permissions`
625- 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
626- Fixed `claude install --force` not bypassing stale lock files
627- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
628- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
629
630## 2.0.65
631
632- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
633- Added context window information to status line input
634- Added `fileSuggestion` setting for custom `@` file search commands
635- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
636- Fixed prompt not being saved to history when aborting a query with Escape
637- Fixed Read tool image handling to identify format from bytes instead of file extension
638
639## 2.0.64
640
641- Made auto-compacting instant
642- Agents and bash commands can run asynchronously and send messages to wake up the main agent
643- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
644- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
645- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
646- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
647- Fixed auto-loading .env when using native installer
648- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
649- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
650- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
651- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
652- Bedrock: Improve efficiency of token counting
653- Bedrock: Add support for `aws login` AWS Management Console credentials
654- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
655
656## 2.0.62
657
658- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
659- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
660- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
661- Fixed slash command selection not working when multiple commands share the same name
662- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
663- Fixed running versions getting removed because lock file incorrectly going stale
664- Fixed IDE diff tab not closing when rejecting file changes
665
666## 2.0.61
667
668- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
669
670## 2.0.60
671
672- Added background agent support. Agents run in the background while you work
673- Added --disable-slash-commands CLI flag to disable all slash commands
674- Added model name to "Co-Authored-By" commit messages
675- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
676- Updated Fetch to skip summarization for pre-approved websites
677- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
678
679## 2.0.59
680
681- Added --agent CLI flag to override the agent setting for the current session
682- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
683- VS Code: Fixed .claude.json config file being read from incorrect location
684
685## 2.0.58
686
687- Pro users now have access to Opus 4.5 as part of their subscription!
688- Fixed timer duration showing "11m 60s" instead of "12m 0s"
689- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
690
691## 2.0.57
692
693- Added feedback input when rejecting plans, allowing users to tell Claude what to change
694- VSCode: Added streaming message support for real-time response display
695
696## 2.0.56
697
698- Added setting to enable/disable terminal progress bar (OSC 9;4)
699- 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.
700
701## 2.0.55
702
703- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
704- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
705- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
706- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
707
708## 2.0.54
709
710- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
711- Fix issue with excessive iTerm notifications
712
713## 2.0.52
714
715- Fixed duplicate message display when starting Claude with a command line argument
716- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
717- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
718- Permit some uses of `$!` in bash commands
719
720## 2.0.51
721
722- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
723- Introducing Claude Code for Desktop: https://claude.com/download
724- 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
725- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
726- Plan Mode now builds more precise plans and executes more thoroughly
727- Usage limit notifications now easier to understand
728- Switched `/usage` back to "% used"
729- Fixed handling of thinking errors
730- Fixed performance regression
731
732## 2.0.50
733
734- Fixed bug preventing calling MCP tools that have nested references in their input schemas
735- Silenced a noisy but harmless error during upgrades
736- Improved ultrathink text display
737- Improved clarity of 5-hour session limit warning message
738
739## 2.0.49
740
741- Added readline-style ctrl-y for pasting deleted text
742- Improved clarity of usage limit warning message
743- Fixed handling of subagent permissions
744
745## 2.0.47
746
747- Improved error messages and validation for `claude --teleport`
748- Improved error handling in `/usage`
749- Fixed race condition with history entry not getting logged at exit
750- Fixed Vertex AI configuration not being applied from `settings.json`
751
752## 2.0.46
753
754- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
755
756## 2.0.45
757
758- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
759- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
760- Send background tasks to Claude Code on the web by starting a message with `&`
761
762## 2.0.43
763
764- Added `permissionMode` field for custom agents
765- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
766- Added skills frontmatter field to declare skills to auto-load for subagents
767- Added the `SubagentStart` hook event
768- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
769- Fixed duplicate rendering of some messages in the UI
770- Fixed some visual flickers
771- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
772
773## 2.0.42
774
775- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
776
777## 2.0.41
778
779- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
780- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
781- Fixed incorrect labeling of user settings vs project settings in command descriptions
782- Fixed crash when plugin command hooks timeout during execution
783- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
784- Fixed broken security documentation links in trust dialogs and onboarding
785- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
786- ctrl-r history search landing on a slash command no longer cancels the search
787- SDK: Support custom timeouts for hooks
788- Allow more safe git commands to run without approval
789- Plugins: Added support for sharing and installing output styles
790- Teleporting a session from web will automatically set the upstream branch
791
792## 2.0.37
793
794- Fixed how idleness is computed for notifications
795- Hooks: Added matcher values for Notification hook events
796- Output Styles: Added `keep-coding-instructions` option to frontmatter
797
798## 2.0.36
799
800- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
801- Fixed queued messages being incorrectly executed as bash commands
802- Fixed input being lost when typing while a queued message is processed
803
804## 2.0.35
805
806- Improve fuzzy search results when searching commands
807- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
808- 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
809- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
810- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
811
812## 2.0.34
813
814- VSCode Extension: Added setting to configure the initial permission mode for new conversations
815- Improved file path suggestion performance with native Rust-based fuzzy finder
816- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
817- Fixed memory crash when reading or writing large files (especially base64-encoded images)
818
819## 2.0.33
820
821- Native binary installs now launch quicker.
822- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
823- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
824
825## 2.0.32
826
827- Un-deprecate output styles based on community feedback
828- Added `companyAnnouncements` setting for displaying announcements on startup
829- Fixed hook progress messages not updating correctly during PostToolUse hook execution
830
831## 2.0.31
832
833- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
834- Vertex: add support for Web Search on supported models
835- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
836- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
837- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
838- Fixed plugin uninstall not removing plugins
839
840## 2.0.30
841
842- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
843- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
844- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
845- Added prompt-based stop hooks
846- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
847- Enabled SSE MCP servers on native build
848- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
849- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
850- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
851- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
852- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
853- Fixed bug that saved session permissions to local settings
854- Fixed MCP tools not being available to sub-agents
855- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
856- Fixed delay when navigating through typeahead suggestions with arrow keys
857- VSCode: Restored selection indicator in input footer showing current file or code selection status
858
859## 2.0.28
860
861- Plan mode: introduced new Plan subagent
862- Subagents: claude can now choose to resume subagents
863- Subagents: claude can dynamically choose the model used by its subagents
864- SDK: added --max-budget-usd flag
865- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
866- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
867- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
868- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
869- Various other bug fixes
870
871## 2.0.27
872
873- New UI for permission prompts
874- Added current branch filtering and search to session resume screen for easier navigation
875- Fixed directory @-mention causing "No assistant message found" error
876- VSCode Extension: Add config setting to include .gitignored files in file searches
877- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
878
879## 2.0.25
880
881- 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
882
883## 2.0.24
884
885- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
886- Claude Code Web: Support for Web -> CLI teleport
887- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
888- Bedrock: Display awsAuthRefresh output when auth is required
889
890## 2.0.22
891
892- Fixed content layout shift when scrolling through slash commands
893- IDE: Add toggle to enable/disable thinking.
894- Fix bug causing duplicate permission prompts with parallel tool calls
895- Add support for enterprise managed MCP allowlist and denylist
896
897## 2.0.21
898
899- Support MCP `structuredContent` field in tool responses
900- Added an interactive question tool
901- Claude will now ask you questions more often in plan mode
902- Added Haiku 4.5 as a model option for Pro users
903- Fixed an issue where queued commands don't have access to previous messages' output
904
905## 2.0.20
906
907- Added support for Claude Skills
908
909## 2.0.19
910
911- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
912- Fixed a bug where Haiku was unnecessarily called in print mode
913
914## 2.0.17
915
916- Added Haiku 4.5 to model selector!
917- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
918- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
919- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
920- OTEL: support HTTP_PROXY and HTTPS_PROXY
921- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
922
923## 2.0.15
924
925- Fixed bug with resuming where previously created files needed to be read again before writing
926- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
927
928## 2.0.14
929
930- Fix @-mentioning MCP servers to toggle them on/off
931- Improve permission checks for bash with inline env vars
932- Fix ultrathink + thinking toggle
933- Reduce unnecessary logins
934- Document --system-prompt
935- Several improvements to rendering
936- Plugins UI polish
937
938## 2.0.13
939
940- Fixed `/plugin` not working on native build
941
942## 2.0.12
943
944- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
945- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
946- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
947- `/plugin validate` command for validating plugin structure and configuration
948- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
949- Plugin documentation available at https://code.claude.com/docs/en/plugins
950- Comprehensive error messages and diagnostics via `/doctor` command
951- Avoid flickering in `/model` selector
952- Improvements to `/help`
953- Avoid mentioning hooks in `/resume` summaries
954- Changes to the "verbose" setting in `/config` now persist across sessions
955
956## 2.0.11
957
958- Reduced system prompt size by 1.4k tokens
959- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
960- Fixed Opus fallback rate limit errors appearing incorrectly
961- Fixed /add-dir command selecting wrong default tab
962
963## 2.0.10
964
965- Rewrote terminal renderer for buttery smooth UI
966- Enable/disable MCP servers by @mentioning, or in /mcp
967- Added tab completion for shell commands in bash mode
968- PreToolUse hooks can now modify tool inputs
969- Press Ctrl-G to edit your prompt in your system's configured text editor
970- Fixes for bash permission checks with environment variables in the command
971
972## 2.0.9
973
974- Fix regression where bash backgrounding stopped working
975
976## 2.0.8
977
978- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
979- IDE: Add drag-and-drop support for files and folders in chat
980- /context: Fix counting for thinking blocks
981- Improve message rendering for users with light themes on dark terminals
982- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
983
984## 2.0.5
985
986- IDE: Fix IME unintended message submission with Enter and Tab
987- IDE: Add "Open in Terminal" link in login screen
988- Fix unhandled OAuth expiration 401 API errors
989- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
990
991## 2.0.1
992
993- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
994- Various bug fixes and presentation improvements
995
996## 2.0.0
997
998- New native VS Code extension
999- Fresh coat of paint throughout the whole app
1000- /rewind a conversation to undo code changes
1001- /usage command to see plan limits
1002- Tab to toggle thinking (sticky across sessions)
1003- Ctrl-R to search history
1004- Unshipped claude config command
1005- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1006- SDK: The Claude Code SDK is now the Claude Agent SDK
1007- Add subagents dynamically with `--agents` flag
1008
1009## 1.0.126
1010
1011- Enable /context command for Bedrock and Vertex
1012- Add mTLS support for HTTP-based OpenTelemetry exporters
1013
1014## 1.0.124
1015
1016- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1017- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1018- No longer inform Claude of the list of allowed tools when permission is denied
1019- Fixed security vulnerability in Bash tool permission checks
1020- Improved VSCode extension performance for large files
1021
1022## 1.0.123
1023
1024- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1025- Fixed thinking mode triggering on negation phrases like "don't think"
1026- Fixed rendering performance degradation during token streaming
1027- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1028- Enhanced BashTool environment snapshot logging
1029- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1030- Migrated --debug logging to a file, to enable easy tailing & filtering
1031
1032## 1.0.120
1033
1034- Fix input lag during typing, especially noticeable with large prompts
1035- Improved VSCode extension command registry and sessions dialog user experience
1036- Enhanced sessions dialog responsiveness and visual feedback
1037- Fixed IDE compatibility issue by removing worktree support check
1038- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1039
1040## 1.0.119
1041
1042- Fix Windows issue where process visually freezes on entering interactive mode
1043- Support dynamic headers for MCP servers via headersHelper configuration
1044- Fix thinking mode not working in headless sessions
1045- Fix slash commands now properly update allowed tools instead of replacing them
1046
1047## 1.0.117
1048
1049- Add Ctrl-R history search to recall previous commands like bash/zsh
1050- Fix input lag while typing, especially on Windows
1051- Add sed command to auto-allowed commands in acceptEdits mode
1052- Fix Windows PATH comparison to be case-insensitive for drive letters
1053- Add permissions management hint to /add-dir output
1054
1055## 1.0.115
1056
1057- Improve thinking mode display with enhanced visual effects
1058- Type /t to temporarily disable thinking mode in your prompt
1059- Improve path validation for glob and grep tools
1060- Show condensed output for post-tool hooks to reduce visual clutter
1061- Fix visual feedback when loading state completes
1062- Improve UI consistency for permission request dialogs
1063
1064## 1.0.113
1065
1066- Deprecated piped input in interactive mode
1067- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1068
1069## 1.0.112
1070
1071- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1072- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1073- Hooks: Added systemMessage support for SessionEnd hooks
1074- Added `spinnerTipsEnabled` setting to disable spinner tips
1075- IDE: Various improvements and bug fixes
1076
1077## 1.0.111
1078
1079- /model now validates provided model names
1080- Fixed Bash tool crashes caused by malformed shell syntax parsing
1081
1082## 1.0.110
1083
1084- /terminal-setup command now supports WezTerm
1085- MCP: OAuth tokens now proactively refresh before expiration
1086- Fixed reliability issues with background Bash processes
1087
1088## 1.0.109
1089
1090- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1091
1092## 1.0.106
1093
1094- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1095
1096## 1.0.97
1097
1098- Settings: /doctor now validates permission rule syntax and suggests corrections
1099
1100## 1.0.94
1101
1102- Vertex: add support for global endpoints for supported models
1103- /memory command now allows direct editing of all imported memory files
1104- SDK: Add custom tools as callbacks
1105- Added /todos command to list current todo items
1106
1107## 1.0.93
1108
1109- Windows: Add alt + v shortcut for pasting images from clipboard
1110- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1111
1112## 1.0.90
1113
1114- Settings file changes take effect immediately - no restart required
1115
1116## 1.0.88
1117
1118- Fixed issue causing "OAuth authentication is currently not supported"
1119- Status line input now includes `exceeds_200k_tokens`
1120- Fixed incorrect usage tracking in /cost.
1121- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1122- Bedrock: Updated default Sonnet model to Sonnet 4
1123
1124## 1.0.86
1125
1126- Added /context to help users self-serve debug context issues
1127- SDK: Added UUID support for all SDK messages
1128- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1129
1130## 1.0.85
1131
1132- Status line input now includes session cost info
1133- Hooks: Introduced SessionEnd hook
1134
1135## 1.0.84
1136
1137- Fix tool_use/tool_result id mismatch error when network is unstable
1138- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1139- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1140- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1141
1142## 1.0.83
1143
1144- @-mention: Support files with spaces in path
1145- New shimmering spinner
1146
1147## 1.0.82
1148
1149- SDK: Add request cancellation support
1150- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1151- Settings: Validation prevents invalid fields in .claude/settings.json files
1152- MCP: Improve tool name consistency
1153- Bash: Fix crash when Claude tries to automatically read large files
1154
1155## 1.0.81
1156
1157- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1158- Agents: Fix custom agent loading when agent files are unparsable
1159
1160## 1.0.80
1161
1162- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1163
1164## 1.0.77
1165
1166- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1167- SDK: Add session support and permission denial tracking
1168- Fix token limit errors in conversation summarization
1169- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1170
1171## 1.0.73
1172
1173- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1174- MCP: Press Esc to cancel OAuth authentication flows
1175- Bash: Improved command validation and reduced false security warnings
1176- UI: Enhanced spinner animations and status line visual hierarchy
1177- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1178
1179## 1.0.72
1180
1181- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1182
1183## 1.0.71
1184
1185- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1186- Customizable status line: add your terminal prompt to Claude Code with /statusline
1187
1188## 1.0.70
1189
1190- Performance: Optimized message rendering for better performance with large contexts
1191- Windows: Fixed native file search, ripgrep, and subagent functionality
1192- Added support for @-mentions in slash command arguments
1193
1194## 1.0.69
1195
1196- Upgraded Opus to version 4.1
1197
1198## 1.0.68
1199
1200- Fix incorrect model names being used for certain commands like `/pr-comments`
1201- 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.
1202- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1203- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1204- SDK: Added canUseTool callback support for tool confirmation
1205- Added `disableAllHooks` setting
1206- Improved file suggestions performance in large repos
1207
1208## 1.0.65
1209
1210- IDE: Fixed connection stability issues and error handling for diagnostics
1211- Windows: Fixed shell environment setup for users without .bashrc files
1212
1213## 1.0.64
1214
1215- Agents: Added model customization support - you can now specify which model an agent should use
1216- Agents: Fixed unintended access to the recursive agent tool
1217- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1218- SDK: Fixed user input tracking across multi-turn conversations
1219- Added hidden files to file search and @-mention suggestions
1220
1221## 1.0.63
1222
1223- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1224
1225## 1.0.62
1226
1227- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1228- Hooks: Added SessionStart hook for new session initialization
1229- /add-dir command now supports typeahead for directory paths
1230- Improved network connectivity check reliability
1231
1232## 1.0.61
1233
1234- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1235- Settings: Added `--settings` flag to load settings from a JSON file
1236- Settings: Fixed resolution of settings files paths that are symlinks
1237- OTEL: Fixed reporting of wrong organization after authentication changes
1238- Slash commands: Fixed permissions checking for allowed-tools with Bash
1239- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1240- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1241- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1242
1243## 1.0.60
1244
1245- You can now create custom subagents for specialized tasks! Run /agents to get started
1246
1247## 1.0.59
1248
1249- SDK: Added tool confirmation support with canUseTool callback
1250- SDK: Allow specifying env for spawned process
1251- Hooks: Exposed PermissionDecision to hooks (including "ask")
1252- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1253- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1254
1255## 1.0.58
1256
1257- Added support for reading PDFs
1258- MCP: Improved server health status display in 'claude mcp list'
1259- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1260
1261## 1.0.57
1262
1263- Added support for specifying a model in slash commands
1264- Improved permission messages to help Claude understand allowed tools
1265- Fix: Remove trailing newlines from bash output in terminal wrapping
1266
1267## 1.0.56
1268
1269- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1270- Fixes for WSL IDE detection
1271- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1272
1273## 1.0.55
1274
1275- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1276- Windows: fixed Ctrl+Z crash
1277- SDK: Added ability to capture error logging
1278- Add --system-prompt-file option to override system prompt in print mode
1279
1280## 1.0.54
1281
1282- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1283- Custom slash commands: Added argument-hint to frontmatter
1284- Windows: OAuth uses port 45454 and properly constructs browser URL
1285- Windows: mode switching now uses alt + m, and plan mode renders properly
1286- Shell: Switch to in-memory shell snapshot to fix file-related errors
1287
1288## 1.0.53
1289
1290- Updated @-mention file truncation from 100 lines to 2000 lines
1291- 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).
1292
1293## 1.0.52
1294
1295- Added support for MCP server instructions
1296
1297## 1.0.51
1298
1299- Added support for native Windows (requires Git for Windows)
1300- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1301- Settings: /doctor can now help you identify and fix invalid setting files
1302- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1303- Increased auto-compact warning threshold from 60% to 80%
1304- Fixed an issue with handling user directories with spaces for shell snapshots
1305- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1306- Custom slash commands: Fixed user-level commands in subdirectories
1307- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1308
1309## 1.0.48
1310
1311- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1312- Added progress messages to Bash tool based on the last 5 lines of command output
1313- Added expanding variables support for MCP server configuration
1314- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1315- Improved IDE extension path handling when Claude Code runs in WSL
1316- Hooks: Added a PreCompact hook
1317- Vim mode: Added c, f/F, t/T
1318
1319## 1.0.45
1320
1321- Redesigned Search (Grep) tool with new tool input parameters and features
1322- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1323- Fixed config file corruption issue by enforcing atomic writes
1324- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1325- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1326- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1327
1328## 1.0.44
1329
1330- New /export command lets you quickly export a conversation for sharing
1331- MCP: resource_link tool results are now supported
1332- MCP: tool annotations and tool titles now display in /mcp view
1333- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1334
1335## 1.0.43
1336
1337- Fixed a bug where the theme selector was saving excessively
1338- Hooks: Added EPIPE system error handling
1339
1340## 1.0.42
1341
1342- Added tilde (`~`) expansion support to `/add-dir` command
1343
1344## 1.0.41
1345
1346- Hooks: Split Stop hook triggering into Stop and SubagentStop
1347- Hooks: Enabled optional timeout configuration for each command
1348- Hooks: Added "hook_event_name" to hook input
1349- Fixed a bug where MCP tools would display twice in tool list
1350- New tool parameters JSON for Bash tool in `tool_decision` event
1351
1352## 1.0.40
1353
1354- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1355
1356## 1.0.39
1357
1358- New Active Time metric in OpenTelemetry logging
1359
1360## 1.0.38
1361
1362- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1363
1364## 1.0.37
1365
1366- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1367
1368## 1.0.36
1369
1370- Web search now takes today's date into context
1371- Fixed a bug where stdio MCP servers were not terminating properly on exit
1372
1373## 1.0.35
1374
1375- Added support for MCP OAuth Authorization Server discovery
1376
1377## 1.0.34
1378
1379- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1380
1381## 1.0.33
1382
1383- Improved logging functionality with session ID support
1384- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1385- Improvements to plan mode
1386
1387## 1.0.32
1388
1389- Updated loopback config for litellm
1390- Added forceLoginMethod setting to bypass login selection screen
1391
1392## 1.0.31
1393
1394- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1395
1396## 1.0.30
1397
1398- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1399- Improved file path autocomplete with filename matching
1400- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1401- Enhanced jq regex support for complex filters with pipes and select
1402
1403## 1.0.29
1404
1405- Improved CJK character support in cursor navigation and rendering
1406
1407## 1.0.28
1408
1409- Slash commands: Fix selector display during history navigation
1410- Resizes images before upload to prevent API size limit errors
1411- Added XDG_CONFIG_HOME support to configuration directory
1412- Performance optimizations for memory usage
1413- New attributes (terminal.type, language) in OpenTelemetry logging
1414
1415## 1.0.27
1416
1417- Streamable HTTP MCP servers are now supported
1418- Remote MCP servers (SSE and HTTP) now support OAuth
1419- MCP resources can now be @-mentioned
1420- /resume slash command to switch conversations within Claude Code
1421
1422## 1.0.25
1423
1424- Slash commands: moved "project" and "user" prefixes to descriptions
1425- Slash commands: improved reliability for command discovery
1426- Improved support for Ghostty
1427- Improved web search reliability
1428
1429## 1.0.24
1430
1431- Improved /mcp output
1432- Fixed a bug where settings arrays got overwritten instead of merged
1433
1434## 1.0.23
1435
1436- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1437- Released Python SDK: pip install claude-code-sdk to get started
1438
1439## 1.0.22
1440
1441- SDK: Renamed `total_cost` to `total_cost_usd`
1442
1443## 1.0.21
1444
1445- Improved editing of files with tab-based indentation
1446- Fix for tool_use without matching tool_result errors
1447- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1448
1449## 1.0.18
1450
1451- Added --add-dir CLI argument for specifying additional working directories
1452- Added streaming input support without require -p flag
1453- Improved startup performance and session storage performance
1454- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1455- Added detailed MCP server tools display (/mcp)
1456- MCP authentication and permission improvements
1457- Added auto-reconnection for MCP SSE connections on disconnect
1458- Fixed issue where pasted content was lost when dialogs appeared
1459
1460## 1.0.17
1461
1462- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1463- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1464- MCP server list UI improvements
1465- Update Claude Code process title to display "claude" instead of "node"
1466
1467## 1.0.11
1468
1469- Claude Code can now also be used with a Claude Pro subscription
1470- Added /upgrade for smoother switching to Claude Max plans
1471- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1472- Improved shell configuration error handling
1473- Improved todo list handling during compaction
1474
1475## 1.0.10
1476
1477- Added markdown table support
1478- Improved streaming performance
1479
1480## 1.0.8
1481
1482- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1483- Increased default otel interval from 1s -> 5s
1484- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1485- Fixed a regression where search tools unnecessarily asked for permissions
1486- Added support for triggering thinking non-English languages
1487- Improved compacting UI
1488
1489## 1.0.7
1490
1491- Renamed /allowed-tools -> /permissions
1492- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1493- Deprecated claude config commands in favor of editing settings.json
1494- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1495- Improved error handling for /install-github-app
1496- Bugfixes, UI polish, and tool reliability improvements
1497
1498## 1.0.6
1499
1500- Improved edit reliability for tab-indented files
1501- Respect CLAUDE_CONFIG_DIR everywhere
1502- Reduced unnecessary tool permission prompts
1503- Added support for symlinks in @file typeahead
1504- Bugfixes, UI polish, and tool reliability improvements
1505
1506## 1.0.4
1507
1508- Fixed a bug where MCP tool errors weren't being parsed correctly
1509
1510## 1.0.1
1511
1512- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1513- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1514- Updated documentation links and OAuth process descriptions
1515
1516## 1.0.0
1517
1518- Claude Code is now generally available
1519- Introducing Sonnet 4 and Opus 4 models
1520
1521## 0.2.125
1522
1523- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1524- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1525
1526## 0.2.117
1527
1528- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1529- Introduced settings.cleanupPeriodDays
1530- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1531- Introduced --debug mode
1532
1533## 0.2.108
1534
1535- You can now send messages to Claude while it works to steer Claude in real-time
1536- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1537- Fixed a bug where thinking was not working in -p mode
1538- Fixed a regression in /cost reporting
1539- Deprecated MCP wizard interface in favor of other MCP commands
1540- Lots of other bugfixes and improvements
1541
1542## 0.2.107
1543
1544- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1545
1546## 0.2.106
1547
1548- MCP SSE server configs can now specify custom headers
1549- Fixed a bug where MCP permission prompt didn't always show correctly
1550
1551## 0.2.105
1552
1553- Claude can now search the web
1554- Moved system & account status to /status
1555- Added word movement keybindings for Vim
1556- Improved latency for startup, todo tool, and file edits
1557
1558## 0.2.102
1559
1560- Improved thinking triggering reliability
1561- Improved @mention reliability for images and folders
1562- You can now paste multiple large chunks into one prompt
1563
1564## 0.2.100
1565
1566- Fixed a crash caused by a stack overflow error
1567- Made db storage optional; missing db support disables --continue and --resume
1568
1569## 0.2.98
1570
1571- Fixed an issue where auto-compact was running twice
1572
1573## 0.2.96
1574
1575- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1576
1577## 0.2.93
1578
1579- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1580- Claude now has access to a Todo list that helps it stay on track and be more organized
1581
1582## 0.2.82
1583
1584- Added support for --disallowedTools
1585- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1586
1587## 0.2.75
1588
1589- Hit Enter to queue up additional messages while Claude is working
1590- Drag in or copy/paste image files directly into the prompt
1591- @-mention files to directly add them to context
1592- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1593- Improved performance for filename auto-complete
1594
1595## 0.2.74
1596
1597- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1598- Task tool can now perform writes and run bash commands
1599
1600## 0.2.72
1601
1602- Updated spinner to indicate tokens loaded and tool usage
1603
1604## 0.2.70
1605
1606- Network commands like curl are now available for Claude to use
1607- Claude can now run multiple web queries in parallel
1608- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1609
1610## 0.2.69
1611
1612- Fixed UI glitches with improved Select component behavior
1613- Enhanced terminal output display with better text truncation logic
1614
1615## 0.2.67
1616
1617- Shared project permission rules can be saved in .claude/settings.json
1618
1619## 0.2.66
1620
1621- Print mode (-p) now supports streaming output via --output-format=stream-json
1622- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1623
1624## 0.2.63
1625
1626- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1627
1628## 0.2.61
1629
1630- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1631- Enhanced image detection for more reliable clipboard paste functionality
1632- Fixed an issue where ESC key could crash the conversation history selector
1633
1634## 0.2.59
1635
1636- Copy+paste images directly into your prompt
1637- Improved progress indicators for bash and fetch tools
1638- Bugfixes for non-interactive mode (-p)
1639
1640## 0.2.54
1641
1642- Quickly add to Memory by starting your message with '#'
1643- Press ctrl+r to see full output for long tool results
1644- Added support for MCP SSE transport
1645
1646## 0.2.53
1647
1648- New web fetch tool lets Claude view URLs that you paste in
1649- Fixed a bug with JPEG detection
1650
1651## 0.2.50
1652
1653- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1654
1655## 0.2.49
1656
1657- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1658
1659## 0.2.47
1660
1661- Press Tab to auto-complete file and folder names
1662- Press Shift + Tab to toggle auto-accept for file edits
1663- Automatic conversation compaction for infinite conversation length (toggle with /config)
1664
1665## 0.2.44
1666
1667- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1668
1669## 0.2.41
1670
1671- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1672- MCP server startup no longer blocks the app from starting up
1673
1674## 0.2.37
1675
1676- New /release-notes command lets you view release notes at any time
1677- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1678
1679## 0.2.36
1680
1681- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1682- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1683
1684## 0.2.34
1685
1686- Vim bindings for text input - enable with /vim or /config
1687
1688## 0.2.32
1689
1690- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1691- Fix for some PersistentShell issues
1692
1693## 0.2.31
1694
1695- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1696- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1697
1698## 0.2.30
1699
1700- Added ANSI color theme for better terminal compatibility
1701- Fixed issue where slash command arguments weren't being sent properly
1702- (Mac-only) API keys are now stored in macOS Keychain
1703
1704## 0.2.26
1705
1706- New /approved-tools command for managing tool permissions
1707- Word-level diff display for improved code readability
1708- Fuzzy matching for slash commands
1709
1710## 0.2.21
1711
1712- Fuzzy matching for /commands