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 d787369 · on Feb 16, 2026 History
1,711 lines · 90.0 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.44
4
5- Fixed auth refresh errors
6
7## 2.1.43
8
9- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
10- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
11- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
12
13## 2.1.42
14
15- Improved startup performance by deferring Zod schema construction
16- Improved prompt cache hit rates by moving date out of system prompt
17- Added one-time Opus 4.6 effort callout for eligible users
18- Fixed /resume showing interrupt messages as session titles
19- Fixed image dimension limit errors to suggest /compact
20
21## 2.1.41
22
23- Added guard against launching Claude Code inside another Claude Code session
24- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
25- Fixed a crash when MCP tools return image content during streaming
26- Fixed /resume session previews showing raw XML tags instead of readable command names
27- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
28- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
29- Fixed hook blocking errors (exit code 2) not showing stderr to the user
30- Added `speed` attribute to OTel events and trace spans for fast mode visibility
31- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
32- Added Windows ARM64 (win32-arm64) native binary support
33- Improved `/rename` to auto-generate session name from conversation context when called without arguments
34- Improved narrow terminal layout for prompt footer
35- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
36- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
37- Fixed background task notifications not being delivered in streaming Agent SDK mode
38- Fixed cursor jumping to end on each keystroke in classifier rule input
39- Fixed markdown link display text being dropped for raw URL
40- Fixed auto-compact failure error notifications being shown to users
41- Fixed permission wait time being included in subagent elapsed time display
42- Fixed proactive ticks firing while in plan mode
43- Fixed clear stale permission rules when settings change on disk
44- Fixed hook blocking errors showing stderr content in UI
45
46## 2.1.39
47
48- Improved terminal rendering performance
49- Fixed fatal errors being swallowed instead of displayed
50- Fixed process hanging after session close
51- Fixed character loss at terminal screen boundary
52- Fixed blank lines in verbose transcript view
53
54## 2.1.38
55
56- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
57- Fixed Tab key queueing slash commands instead of autocompleting
58- Fixed bash permission matching for commands using environment variable wrappers
59- Fixed text between tool uses disappearing when not using streaming
60- Fixed duplicate sessions when resuming in VS Code extension
61- Improved heredoc delimiter parsing to prevent command smuggling
62- Blocked writes to `.claude/skills` directory in sandbox mode
63
64## 2.1.37
65
66- Fixed an issue where /fast was not immediately available after enabling /extra-usage
67
68## 2.1.36
69
70- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
71
72## 2.1.34
73
74- Fixed a crash when agent teams setting changed between renders
75- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
76
77## 2.1.33
78
79- Fixed agent teammate sessions in tmux to send and receive messages
80- Fixed warnings about agent teams not being available on your current plan
81- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
82- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
83- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
84- Added plugin name to skill descriptions and `/skills` menu for better discoverability
85- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
86- 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
87- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
88- 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
89- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
90- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
91- Errors from invalid managed settings are now surfaced
92- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
93- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
94- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
95
96## 2.1.32
97
98- Claude Opus 4.6 is now available!
99- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
100- Claude now automatically records and recalls memories as it works
101- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
102- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
103- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
104- Updated --resume to re-use --agent value specified in previous conversation by default.
105- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
106- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
107- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
108- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
109- VSCode: Added spinner when loading past conversations list
110
111## 2.1.31
112
113- Added session resume hint on exit, showing how to continue your conversation later
114- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
115- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
116- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
117- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
118- 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
119- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
120- 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
121- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
122- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
123- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
124
125## 2.1.30
126
127- 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.
128- 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`.
129- Added `/debug` for Claude to help troubleshoot the current session
130- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
131- Added token count, tool uses, and duration metrics to Task tool results
132- Added reduced motion mode to the config
133- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
134- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
135- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
136- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
137- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
138- Fixed permission dialogs stealing focus while actively typing
139- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
140- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
141- 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
142- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
143- Changed `/model` to execute immediately instead of being queued
144- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
145- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
146
147## 2.1.29
148
149- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
150
151## 2.1.27
152
153- Added tool call failures and denials to debug logs
154- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
155- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
156- Sessions are now automatically linked to PRs when created via `gh pr create`
157- Fixed /context command not displaying colored output
158- Fixed status bar duplicating background task indicator when PR status was shown
159- Windows: Fixed bash command execution failing for users with `.bashrc` files
160- Windows: Fixed console windows flashing when spawning child processes
161- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
162
163## 2.1.25
164
165- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
166
167## 2.1.23
168
169- Added customizable spinner verbs setting (`spinnerVerbs`)
170- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
171- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
172- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
173- Fixed pending async hooks not being cancelled when headless streaming sessions ended
174- Fixed tab completion not updating the input field when accepting a suggestion
175- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
176- Improved terminal rendering performance with optimized screen data layout
177- Changed Bash commands to show timeout duration alongside elapsed time
178- Changed merged pull requests to show a purple status indicator in the prompt footer
179- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
180
181## 2.1.22
182
183- Fixed structured outputs for non-interactive (-p) mode
184
185## 2.1.21
186
187- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
188- Fixed shell completion cache files being truncated on exit
189- Fixed API errors when resuming sessions that were interrupted during tool execution
190- Fixed auto-compact triggering too early on models with large output token limits
191- Fixed task IDs potentially being reused after deletion
192- Fixed file search not working in VS Code extension on Windows
193- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
194- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
195- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
196- [VSCode] Fixed message action buttons having incorrect background colors
197
198## 2.1.20
199
200- Added arrow key history navigation in vim normal mode when cursor cannot move further
201- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
202- 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
203- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
204- Added ability to delete tasks via the `TaskUpdate` tool
205- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
206- Fixed agents sometimes ignoring user messages sent while actively working on a task
207- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
208- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
209- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
210- Fixed draft prompt being lost when pressing UP arrow to navigate command history
211- Fixed ghost text flickering when typing slash commands mid-input
212- Fixed marketplace source removal not properly deleting settings
213- Fixed duplicate output in some commands like `/context`
214- Fixed task list sometimes showing outside the main conversation view
215- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
216- Fixed crashes when cancelling tool use
217- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
218- Improved thinking status text with a subtle shimmer animation
219- Improved task list to dynamically adjust visible items based on terminal height
220- Improved fork conversation hint to show how to resume the original session
221- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
222- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
223- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
224- Changed the `/copy` command to be available to all users
225- Changed background agents to prompt for tool permissions before launching
226- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
227- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
228
229## 2.1.19
230
231- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
232- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
233- Fixed crashes on processors without AVX instruction support
234- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
235- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
236- Fixed resuming sessions by custom title not working when run from a different directory
237- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
238- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
239- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
240- Fixed file write preview omitting empty lines
241- Changed skills without additional permissions or hooks to be allowed without requiring approval
242- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
243- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
244- [VSCode] Enabled session forking and rewind functionality for all users
245
246## 2.1.18
247
248- 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
249
250## 2.1.17
251
252- Fixed crashes on processors without AVX instruction support
253
254## 2.1.16
255
256- Added new task management system, including new capabilities like dependency tracking
257- [VSCode] Added native plugin management support
258- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
259- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
260- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
261- Fixed session titles on the resume screen not respecting the user's language setting
262- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
263
264## 2.1.15
265
266- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
267- Improved UI rendering performance with React Compiler
268- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
269- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
270
271## 2.1.14
272
273- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
274- Added search to installed plugins list - type to filter by name or description
275- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
276- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
277- Fixed memory issues that could cause crashes when running parallel subagents
278- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
279- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
280- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
281- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
282- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
283- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
284- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
285- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
286- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
287- Improved backspace to delete pasted text as a single token instead of one character at a time
288- [VSCode] Added `/usage` command to display current plan usage
289
290## 2.1.12
291
292- Fixed message rendering bug
293
294## 2.1.11
295
296- Fixed excessive MCP connection requests for HTTP/SSE transports
297
298## 2.1.10
299
300- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
301- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
302- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
303- Improved startup to capture keystrokes typed before the REPL is fully ready
304- Improved file suggestions to show as removable attachments instead of inserting text when accepted
305- [VSCode] Added install count display to plugin listings
306- [VSCode] Added trust warning when installing plugins
307
308## 2.1.9
309
310- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
311- Added `plansDirectory` setting to customize where plan files are stored
312- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
313- Added session URL attribution to commits and PRs created from web sessions
314- Added support for `PreToolUse` hooks to return `additionalContext` to the model
315- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
316- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
317- Fixed MCP server reconnection hanging when cached connection promise never resolves
318- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
319
320## 2.1.7
321
322- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
323- Added ability to provide feedback when accepting permission prompts
324- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
325- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
326- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
327- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
328- 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)
329- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
330- Fixed terminal title animation jitter by using fixed-width braille characters
331- Fixed plugins with git submodules not being fully initialized when installed
332- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
333- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
334- 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.
335- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
336- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
337
338## 2.1.6
339
340- Added search functionality to `/config` command for quickly filtering settings
341- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
342- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
343- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
344- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
345- Added an error display when the editor fails during Ctrl+G
346- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
347- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
348- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
349- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
350- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
351- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
352- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
353- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
354- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
355- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
356- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
357- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
358- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
359- Improved @ autocomplete with icons for different suggestion types and single-line formatting
360- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
361- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
362- Changed terminal title to "Claude Code" on startup for better window identification
363- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
364- [VSCode] Fixed usage indicator not updating after manual compact
365
366## 2.1.5
367
368- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
369
370## 2.1.4
371
372- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
373- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
374
375## 2.1.3
376
377- Merged slash commands and skills, simplifying the mental model with no change in behavior
378- Added release channel (`stable` or `latest`) toggle to `/config`
379- 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
380- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
381- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
382- Fixed mismatch between background task count in status bar and items shown in tasks dialog
383- Fixed sub-agents using the wrong model during conversation compaction
384- Fixed web search in sub-agents using incorrect model
385- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
386- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
387- Improved slash command suggestion readability by truncating long descriptions to 2 lines
388- Changed tool hook execution timeout from 60 seconds to 10 minutes
389- [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)
390
391## 2.1.2
392
393- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
394- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
395- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
396- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
397- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
398- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
399- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
400- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
401- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
402- Fixed updates incorrectly claiming another installation is in progress
403- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
404- Fixed remote session URL and teleport being broken when using `/tasks` command
405- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
406- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
407- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
408- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
409- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
410- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
411- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
412- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
413- [SDK] Changed minimum zod peer dependency to ^4.0.0
414- [VSCode] Fixed usage display not updating after manual compact
415
416## 2.1.0
417
418- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
419- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
420- Added support for `agent` field in skills to specify agent type for execution
421- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
422- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
423- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
424- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
425- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
426- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
427- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
428- Fixed Esc key with queued prompts to only move them to input without canceling the running task
429- Reduced permission prompts for complex bash commands
430- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
431- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
432- Fixed plugin path resolution for file-based marketplace sources
433- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
434- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
435- Fixed Claude in Chrome support for WSL environments
436- Fixed Windows native installer silently failing when executable creation fails
437- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
438- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
439- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
440- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
441- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
442- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
443- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
444- Added hooks support for skill and slash command frontmatter
445- 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
446- Added `/plan` command shortcut to enable plan mode directly from the prompt
447- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
448- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
449- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
450- Added support for `once: true` config for hooks
451- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
452- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
453- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
454- Added left/right arrow key navigation for cycling through tabs in dialogs
455- Added real-time thinking block display in Ctrl+O transcript mode
456- Added filepath to full output in background bash task details dialog
457- Added Skills as a separate category in the context visualization
458- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
459- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
460- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
461- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
462- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
463- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
464- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
465- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
466- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
467- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
468- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
469- Fixed commands with `$()` command substitution failing with parse errors
470- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
471- 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)
472- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
473- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
474- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
475- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
476- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
477- Fixed subagents sometimes not inheriting the parent's model by default
478- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
479- Fixed duplicate Bash commands appearing in permission request option labels
480- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
481- Fixed background task completion notifications to appear proactively with bullet point
482- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
483- Fixed cursor disappearing after dismissing permission dialogs
484- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
485- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
486- Fixed images being silently dropped when queueing messages while backgrounding a task
487- Fixed large pasted images failing with "Image was too large" error
488- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
489- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
490- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
491- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
492- Fixed spinner token counter to properly accumulate tokens from subagents during execution
493- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
494- Fixed race condition where LSP tool could return "no server available" during startup
495- Fixed feedback submission hanging indefinitely when network requests timeout
496- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
497- Fixed hook success message showing trailing colon when hook has no output
498- Multiple optimizations to improve startup performance
499- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
500- Improved performance when reading Jupyter notebooks with many cells
501- Improved reliability for piped input like `cat refactor.md | claude`
502- Improved reliability for AskQuestion tool
503- Improved sed in-place edit commands to render as file edits with diff preview
504- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
505- Improved compaction reliability
506- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
507- Improved skills to show progress while executing, displaying tool uses as they happen
508- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
509- Improved skill suggestions to prioritize recently and frequently used skills
510- Improved spinner feedback when waiting for the first response token
511- Improved token count display in spinner to include tokens from background agents
512- Improved incremental output for async agents to give the main thread more control and visibility
513- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
514- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
515- Improved macOS screenshot paste reliability with TIFF format support
516- Improved `/stats` output
517- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
518- Changed "Interrupted" message color from red to grey for a less alarming appearance
519- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
520- Removed underline styling from image reference links
521- [SDK] Changed minimum zod peer dependency to ^4.0.0
522- [VSCode] Added currently selected model name to the context menu
523- [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")
524- [VSCode] Fixed paragraph breaks not rendering in markdown content
525- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
526- [Windows] Fixed issue with improper rendering
527
528## 2.0.76
529
530- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
531
532## 2.0.75
533
534- Minor bugfixes
535
536## 2.0.74
537
538- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
539- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
540- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
541- Added syntax highlighting info to theme picker
542- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
543- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
544- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
545- Fixed a potential crash when syntax highlighting isn't initialized correctly
546- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
547- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
548- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
549- [Windows] Fixed issue with improper rendering
550- [VSCode] Added gift tag pictogram for year-end promotion message
551
552## 2.0.73
553
554- Added clickable `[Image #N]` links that open attached images in the default viewer
555- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
556- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
557- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
558- Fixed slow input history cycling and race condition that could overwrite text after message submission
559- Improved `/theme` command to open theme picker directly
560- Improved theme picker UI
561- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
562- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
563
564## 2.0.72
565
566- 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
567- Reduced terminal flickering
568- Added scannable QR code to mobile app tip for quick app downloads
569- Added loading indicator when resuming conversations for better feedback
570- Fixed `/context` command not respecting custom system prompts in non-interactive mode
571- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
572- Improved @ mention file suggestion speed (~3× faster in git repositories)
573- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
574- Improved settings validation errors to be more prominent
575- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
576
577## 2.0.71
578
579- Added /config toggle to enable/disable prompt suggestions
580- Added `/settings` as an alias for the `/config` command
581- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
582- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
583- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
584- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
585- New syntax highlighting engine for native build
586
587## 2.0.70
588
589- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
590- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
591- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
592- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
593- Fixed input being cleared when processing queued commands while the user was typing
594- Fixed prompt suggestions replacing typed input when pressing Tab
595- Fixed diff view not updating when terminal is resized
596- Improved memory usage by 3x for large conversations
597- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
598- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
599- Fix thinking mode toggle in /config not persisting correctly
600- Improve UI for file creation permission dialog
601
602## 2.0.69
603
604- Minor bugfixes
605
606## 2.0.68
607
608- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
609- Fixed a bug where disallowed MCP tools were visible to the model
610- Fixed an issue where steering messages could be lost while a subagent is working
611- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
612- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
613- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
614
615## 2.0.67
616
617- Thinking mode is now enabled by default for Opus 4.5
618- Thinking mode configuration has moved to /config
619- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
620- Show reason why autoupdater is disabled in `/doctor`
621- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
622- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
623- Fixed scroll position resetting after deleting a permission rule in `/permissions`
624- 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
625- Fixed `claude install --force` not bypassing stale lock files
626- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
627- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
628
629## 2.0.65
630
631- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
632- Added context window information to status line input
633- Added `fileSuggestion` setting for custom `@` file search commands
634- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
635- Fixed prompt not being saved to history when aborting a query with Escape
636- Fixed Read tool image handling to identify format from bytes instead of file extension
637
638## 2.0.64
639
640- Made auto-compacting instant
641- Agents and bash commands can run asynchronously and send messages to wake up the main agent
642- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
643- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
644- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
645- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
646- Fixed auto-loading .env when using native installer
647- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
648- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
649- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
650- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
651- Bedrock: Improve efficiency of token counting
652- Bedrock: Add support for `aws login` AWS Management Console credentials
653- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
654
655## 2.0.62
656
657- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
658- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
659- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
660- Fixed slash command selection not working when multiple commands share the same name
661- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
662- Fixed running versions getting removed because lock file incorrectly going stale
663- Fixed IDE diff tab not closing when rejecting file changes
664
665## 2.0.61
666
667- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
668
669## 2.0.60
670
671- Added background agent support. Agents run in the background while you work
672- Added --disable-slash-commands CLI flag to disable all slash commands
673- Added model name to "Co-Authored-By" commit messages
674- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
675- Updated Fetch to skip summarization for pre-approved websites
676- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
677
678## 2.0.59
679
680- Added --agent CLI flag to override the agent setting for the current session
681- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
682- VS Code: Fixed .claude.json config file being read from incorrect location
683
684## 2.0.58
685
686- Pro users now have access to Opus 4.5 as part of their subscription!
687- Fixed timer duration showing "11m 60s" instead of "12m 0s"
688- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
689
690## 2.0.57
691
692- Added feedback input when rejecting plans, allowing users to tell Claude what to change
693- VSCode: Added streaming message support for real-time response display
694
695## 2.0.56
696
697- Added setting to enable/disable terminal progress bar (OSC 9;4)
698- 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.
699
700## 2.0.55
701
702- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
703- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
704- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
705- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
706
707## 2.0.54
708
709- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
710- Fix issue with excessive iTerm notifications
711
712## 2.0.52
713
714- Fixed duplicate message display when starting Claude with a command line argument
715- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
716- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
717- Permit some uses of `$!` in bash commands
718
719## 2.0.51
720
721- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
722- Introducing Claude Code for Desktop: https://claude.com/download
723- 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
724- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
725- Plan Mode now builds more precise plans and executes more thoroughly
726- Usage limit notifications now easier to understand
727- Switched `/usage` back to "% used"
728- Fixed handling of thinking errors
729- Fixed performance regression
730
731## 2.0.50
732
733- Fixed bug preventing calling MCP tools that have nested references in their input schemas
734- Silenced a noisy but harmless error during upgrades
735- Improved ultrathink text display
736- Improved clarity of 5-hour session limit warning message
737
738## 2.0.49
739
740- Added readline-style ctrl-y for pasting deleted text
741- Improved clarity of usage limit warning message
742- Fixed handling of subagent permissions
743
744## 2.0.47
745
746- Improved error messages and validation for `claude --teleport`
747- Improved error handling in `/usage`
748- Fixed race condition with history entry not getting logged at exit
749- Fixed Vertex AI configuration not being applied from `settings.json`
750
751## 2.0.46
752
753- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
754
755## 2.0.45
756
757- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
758- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
759- Send background tasks to Claude Code on the web by starting a message with `&`
760
761## 2.0.43
762
763- Added `permissionMode` field for custom agents
764- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
765- Added skills frontmatter field to declare skills to auto-load for subagents
766- Added the `SubagentStart` hook event
767- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
768- Fixed duplicate rendering of some messages in the UI
769- Fixed some visual flickers
770- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
771
772## 2.0.42
773
774- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
775
776## 2.0.41
777
778- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
779- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
780- Fixed incorrect labeling of user settings vs project settings in command descriptions
781- Fixed crash when plugin command hooks timeout during execution
782- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
783- Fixed broken security documentation links in trust dialogs and onboarding
784- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
785- ctrl-r history search landing on a slash command no longer cancels the search
786- SDK: Support custom timeouts for hooks
787- Allow more safe git commands to run without approval
788- Plugins: Added support for sharing and installing output styles
789- Teleporting a session from web will automatically set the upstream branch
790
791## 2.0.37
792
793- Fixed how idleness is computed for notifications
794- Hooks: Added matcher values for Notification hook events
795- Output Styles: Added `keep-coding-instructions` option to frontmatter
796
797## 2.0.36
798
799- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
800- Fixed queued messages being incorrectly executed as bash commands
801- Fixed input being lost when typing while a queued message is processed
802
803## 2.0.35
804
805- Improve fuzzy search results when searching commands
806- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
807- 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
808- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
809- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
810
811## 2.0.34
812
813- VSCode Extension: Added setting to configure the initial permission mode for new conversations
814- Improved file path suggestion performance with native Rust-based fuzzy finder
815- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
816- Fixed memory crash when reading or writing large files (especially base64-encoded images)
817
818## 2.0.33
819
820- Native binary installs now launch quicker.
821- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
822- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
823
824## 2.0.32
825
826- Un-deprecate output styles based on community feedback
827- Added `companyAnnouncements` setting for displaying announcements on startup
828- Fixed hook progress messages not updating correctly during PostToolUse hook execution
829
830## 2.0.31
831
832- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
833- Vertex: add support for Web Search on supported models
834- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
835- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
836- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
837- Fixed plugin uninstall not removing plugins
838
839## 2.0.30
840
841- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
842- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
843- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
844- Added prompt-based stop hooks
845- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
846- Enabled SSE MCP servers on native build
847- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
848- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
849- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
850- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
851- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
852- Fixed bug that saved session permissions to local settings
853- Fixed MCP tools not being available to sub-agents
854- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
855- Fixed delay when navigating through typeahead suggestions with arrow keys
856- VSCode: Restored selection indicator in input footer showing current file or code selection status
857
858## 2.0.28
859
860- Plan mode: introduced new Plan subagent
861- Subagents: claude can now choose to resume subagents
862- Subagents: claude can dynamically choose the model used by its subagents
863- SDK: added --max-budget-usd flag
864- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
865- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
866- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
867- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
868- Various other bug fixes
869
870## 2.0.27
871
872- New UI for permission prompts
873- Added current branch filtering and search to session resume screen for easier navigation
874- Fixed directory @-mention causing "No assistant message found" error
875- VSCode Extension: Add config setting to include .gitignored files in file searches
876- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
877
878## 2.0.25
879
880- 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
881
882## 2.0.24
883
884- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
885- Claude Code Web: Support for Web -> CLI teleport
886- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
887- Bedrock: Display awsAuthRefresh output when auth is required
888
889## 2.0.22
890
891- Fixed content layout shift when scrolling through slash commands
892- IDE: Add toggle to enable/disable thinking.
893- Fix bug causing duplicate permission prompts with parallel tool calls
894- Add support for enterprise managed MCP allowlist and denylist
895
896## 2.0.21
897
898- Support MCP `structuredContent` field in tool responses
899- Added an interactive question tool
900- Claude will now ask you questions more often in plan mode
901- Added Haiku 4.5 as a model option for Pro users
902- Fixed an issue where queued commands don't have access to previous messages' output
903
904## 2.0.20
905
906- Added support for Claude Skills
907
908## 2.0.19
909
910- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
911- Fixed a bug where Haiku was unnecessarily called in print mode
912
913## 2.0.17
914
915- Added Haiku 4.5 to model selector!
916- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
917- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
918- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
919- OTEL: support HTTP_PROXY and HTTPS_PROXY
920- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
921
922## 2.0.15
923
924- Fixed bug with resuming where previously created files needed to be read again before writing
925- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
926
927## 2.0.14
928
929- Fix @-mentioning MCP servers to toggle them on/off
930- Improve permission checks for bash with inline env vars
931- Fix ultrathink + thinking toggle
932- Reduce unnecessary logins
933- Document --system-prompt
934- Several improvements to rendering
935- Plugins UI polish
936
937## 2.0.13
938
939- Fixed `/plugin` not working on native build
940
941## 2.0.12
942
943- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
944- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
945- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
946- `/plugin validate` command for validating plugin structure and configuration
947- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
948- Plugin documentation available at https://code.claude.com/docs/en/plugins
949- Comprehensive error messages and diagnostics via `/doctor` command
950- Avoid flickering in `/model` selector
951- Improvements to `/help`
952- Avoid mentioning hooks in `/resume` summaries
953- Changes to the "verbose" setting in `/config` now persist across sessions
954
955## 2.0.11
956
957- Reduced system prompt size by 1.4k tokens
958- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
959- Fixed Opus fallback rate limit errors appearing incorrectly
960- Fixed /add-dir command selecting wrong default tab
961
962## 2.0.10
963
964- Rewrote terminal renderer for buttery smooth UI
965- Enable/disable MCP servers by @mentioning, or in /mcp
966- Added tab completion for shell commands in bash mode
967- PreToolUse hooks can now modify tool inputs
968- Press Ctrl-G to edit your prompt in your system's configured text editor
969- Fixes for bash permission checks with environment variables in the command
970
971## 2.0.9
972
973- Fix regression where bash backgrounding stopped working
974
975## 2.0.8
976
977- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
978- IDE: Add drag-and-drop support for files and folders in chat
979- /context: Fix counting for thinking blocks
980- Improve message rendering for users with light themes on dark terminals
981- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
982
983## 2.0.5
984
985- IDE: Fix IME unintended message submission with Enter and Tab
986- IDE: Add "Open in Terminal" link in login screen
987- Fix unhandled OAuth expiration 401 API errors
988- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
989
990## 2.0.1
991
992- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
993- Various bug fixes and presentation improvements
994
995## 2.0.0
996
997- New native VS Code extension
998- Fresh coat of paint throughout the whole app
999- /rewind a conversation to undo code changes
1000- /usage command to see plan limits
1001- Tab to toggle thinking (sticky across sessions)
1002- Ctrl-R to search history
1003- Unshipped claude config command
1004- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1005- SDK: The Claude Code SDK is now the Claude Agent SDK
1006- Add subagents dynamically with `--agents` flag
1007
1008## 1.0.126
1009
1010- Enable /context command for Bedrock and Vertex
1011- Add mTLS support for HTTP-based OpenTelemetry exporters
1012
1013## 1.0.124
1014
1015- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1016- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1017- No longer inform Claude of the list of allowed tools when permission is denied
1018- Fixed security vulnerability in Bash tool permission checks
1019- Improved VSCode extension performance for large files
1020
1021## 1.0.123
1022
1023- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1024- Fixed thinking mode triggering on negation phrases like "don't think"
1025- Fixed rendering performance degradation during token streaming
1026- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1027- Enhanced BashTool environment snapshot logging
1028- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1029- Migrated --debug logging to a file, to enable easy tailing & filtering
1030
1031## 1.0.120
1032
1033- Fix input lag during typing, especially noticeable with large prompts
1034- Improved VSCode extension command registry and sessions dialog user experience
1035- Enhanced sessions dialog responsiveness and visual feedback
1036- Fixed IDE compatibility issue by removing worktree support check
1037- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1038
1039## 1.0.119
1040
1041- Fix Windows issue where process visually freezes on entering interactive mode
1042- Support dynamic headers for MCP servers via headersHelper configuration
1043- Fix thinking mode not working in headless sessions
1044- Fix slash commands now properly update allowed tools instead of replacing them
1045
1046## 1.0.117
1047
1048- Add Ctrl-R history search to recall previous commands like bash/zsh
1049- Fix input lag while typing, especially on Windows
1050- Add sed command to auto-allowed commands in acceptEdits mode
1051- Fix Windows PATH comparison to be case-insensitive for drive letters
1052- Add permissions management hint to /add-dir output
1053
1054## 1.0.115
1055
1056- Improve thinking mode display with enhanced visual effects
1057- Type /t to temporarily disable thinking mode in your prompt
1058- Improve path validation for glob and grep tools
1059- Show condensed output for post-tool hooks to reduce visual clutter
1060- Fix visual feedback when loading state completes
1061- Improve UI consistency for permission request dialogs
1062
1063## 1.0.113
1064
1065- Deprecated piped input in interactive mode
1066- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1067
1068## 1.0.112
1069
1070- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1071- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1072- Hooks: Added systemMessage support for SessionEnd hooks
1073- Added `spinnerTipsEnabled` setting to disable spinner tips
1074- IDE: Various improvements and bug fixes
1075
1076## 1.0.111
1077
1078- /model now validates provided model names
1079- Fixed Bash tool crashes caused by malformed shell syntax parsing
1080
1081## 1.0.110
1082
1083- /terminal-setup command now supports WezTerm
1084- MCP: OAuth tokens now proactively refresh before expiration
1085- Fixed reliability issues with background Bash processes
1086
1087## 1.0.109
1088
1089- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1090
1091## 1.0.106
1092
1093- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1094
1095## 1.0.97
1096
1097- Settings: /doctor now validates permission rule syntax and suggests corrections
1098
1099## 1.0.94
1100
1101- Vertex: add support for global endpoints for supported models
1102- /memory command now allows direct editing of all imported memory files
1103- SDK: Add custom tools as callbacks
1104- Added /todos command to list current todo items
1105
1106## 1.0.93
1107
1108- Windows: Add alt + v shortcut for pasting images from clipboard
1109- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1110
1111## 1.0.90
1112
1113- Settings file changes take effect immediately - no restart required
1114
1115## 1.0.88
1116
1117- Fixed issue causing "OAuth authentication is currently not supported"
1118- Status line input now includes `exceeds_200k_tokens`
1119- Fixed incorrect usage tracking in /cost.
1120- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1121- Bedrock: Updated default Sonnet model to Sonnet 4
1122
1123## 1.0.86
1124
1125- Added /context to help users self-serve debug context issues
1126- SDK: Added UUID support for all SDK messages
1127- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1128
1129## 1.0.85
1130
1131- Status line input now includes session cost info
1132- Hooks: Introduced SessionEnd hook
1133
1134## 1.0.84
1135
1136- Fix tool_use/tool_result id mismatch error when network is unstable
1137- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1138- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1139- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1140
1141## 1.0.83
1142
1143- @-mention: Support files with spaces in path
1144- New shimmering spinner
1145
1146## 1.0.82
1147
1148- SDK: Add request cancellation support
1149- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1150- Settings: Validation prevents invalid fields in .claude/settings.json files
1151- MCP: Improve tool name consistency
1152- Bash: Fix crash when Claude tries to automatically read large files
1153
1154## 1.0.81
1155
1156- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1157- Agents: Fix custom agent loading when agent files are unparsable
1158
1159## 1.0.80
1160
1161- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1162
1163## 1.0.77
1164
1165- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1166- SDK: Add session support and permission denial tracking
1167- Fix token limit errors in conversation summarization
1168- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1169
1170## 1.0.73
1171
1172- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1173- MCP: Press Esc to cancel OAuth authentication flows
1174- Bash: Improved command validation and reduced false security warnings
1175- UI: Enhanced spinner animations and status line visual hierarchy
1176- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1177
1178## 1.0.72
1179
1180- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1181
1182## 1.0.71
1183
1184- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1185- Customizable status line: add your terminal prompt to Claude Code with /statusline
1186
1187## 1.0.70
1188
1189- Performance: Optimized message rendering for better performance with large contexts
1190- Windows: Fixed native file search, ripgrep, and subagent functionality
1191- Added support for @-mentions in slash command arguments
1192
1193## 1.0.69
1194
1195- Upgraded Opus to version 4.1
1196
1197## 1.0.68
1198
1199- Fix incorrect model names being used for certain commands like `/pr-comments`
1200- 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.
1201- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1202- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1203- SDK: Added canUseTool callback support for tool confirmation
1204- Added `disableAllHooks` setting
1205- Improved file suggestions performance in large repos
1206
1207## 1.0.65
1208
1209- IDE: Fixed connection stability issues and error handling for diagnostics
1210- Windows: Fixed shell environment setup for users without .bashrc files
1211
1212## 1.0.64
1213
1214- Agents: Added model customization support - you can now specify which model an agent should use
1215- Agents: Fixed unintended access to the recursive agent tool
1216- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1217- SDK: Fixed user input tracking across multi-turn conversations
1218- Added hidden files to file search and @-mention suggestions
1219
1220## 1.0.63
1221
1222- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1223
1224## 1.0.62
1225
1226- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1227- Hooks: Added SessionStart hook for new session initialization
1228- /add-dir command now supports typeahead for directory paths
1229- Improved network connectivity check reliability
1230
1231## 1.0.61
1232
1233- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1234- Settings: Added `--settings` flag to load settings from a JSON file
1235- Settings: Fixed resolution of settings files paths that are symlinks
1236- OTEL: Fixed reporting of wrong organization after authentication changes
1237- Slash commands: Fixed permissions checking for allowed-tools with Bash
1238- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1239- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1240- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1241
1242## 1.0.60
1243
1244- You can now create custom subagents for specialized tasks! Run /agents to get started
1245
1246## 1.0.59
1247
1248- SDK: Added tool confirmation support with canUseTool callback
1249- SDK: Allow specifying env for spawned process
1250- Hooks: Exposed PermissionDecision to hooks (including "ask")
1251- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1252- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1253
1254## 1.0.58
1255
1256- Added support for reading PDFs
1257- MCP: Improved server health status display in 'claude mcp list'
1258- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1259
1260## 1.0.57
1261
1262- Added support for specifying a model in slash commands
1263- Improved permission messages to help Claude understand allowed tools
1264- Fix: Remove trailing newlines from bash output in terminal wrapping
1265
1266## 1.0.56
1267
1268- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1269- Fixes for WSL IDE detection
1270- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1271
1272## 1.0.55
1273
1274- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1275- Windows: fixed Ctrl+Z crash
1276- SDK: Added ability to capture error logging
1277- Add --system-prompt-file option to override system prompt in print mode
1278
1279## 1.0.54
1280
1281- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1282- Custom slash commands: Added argument-hint to frontmatter
1283- Windows: OAuth uses port 45454 and properly constructs browser URL
1284- Windows: mode switching now uses alt + m, and plan mode renders properly
1285- Shell: Switch to in-memory shell snapshot to fix file-related errors
1286
1287## 1.0.53
1288
1289- Updated @-mention file truncation from 100 lines to 2000 lines
1290- 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).
1291
1292## 1.0.52
1293
1294- Added support for MCP server instructions
1295
1296## 1.0.51
1297
1298- Added support for native Windows (requires Git for Windows)
1299- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1300- Settings: /doctor can now help you identify and fix invalid setting files
1301- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1302- Increased auto-compact warning threshold from 60% to 80%
1303- Fixed an issue with handling user directories with spaces for shell snapshots
1304- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1305- Custom slash commands: Fixed user-level commands in subdirectories
1306- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1307
1308## 1.0.48
1309
1310- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1311- Added progress messages to Bash tool based on the last 5 lines of command output
1312- Added expanding variables support for MCP server configuration
1313- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1314- Improved IDE extension path handling when Claude Code runs in WSL
1315- Hooks: Added a PreCompact hook
1316- Vim mode: Added c, f/F, t/T
1317
1318## 1.0.45
1319
1320- Redesigned Search (Grep) tool with new tool input parameters and features
1321- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1322- Fixed config file corruption issue by enforcing atomic writes
1323- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1324- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1325- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1326
1327## 1.0.44
1328
1329- New /export command lets you quickly export a conversation for sharing
1330- MCP: resource_link tool results are now supported
1331- MCP: tool annotations and tool titles now display in /mcp view
1332- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1333
1334## 1.0.43
1335
1336- Fixed a bug where the theme selector was saving excessively
1337- Hooks: Added EPIPE system error handling
1338
1339## 1.0.42
1340
1341- Added tilde (`~`) expansion support to `/add-dir` command
1342
1343## 1.0.41
1344
1345- Hooks: Split Stop hook triggering into Stop and SubagentStop
1346- Hooks: Enabled optional timeout configuration for each command
1347- Hooks: Added "hook_event_name" to hook input
1348- Fixed a bug where MCP tools would display twice in tool list
1349- New tool parameters JSON for Bash tool in `tool_decision` event
1350
1351## 1.0.40
1352
1353- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1354
1355## 1.0.39
1356
1357- New Active Time metric in OpenTelemetry logging
1358
1359## 1.0.38
1360
1361- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1362
1363## 1.0.37
1364
1365- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1366
1367## 1.0.36
1368
1369- Web search now takes today's date into context
1370- Fixed a bug where stdio MCP servers were not terminating properly on exit
1371
1372## 1.0.35
1373
1374- Added support for MCP OAuth Authorization Server discovery
1375
1376## 1.0.34
1377
1378- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1379
1380## 1.0.33
1381
1382- Improved logging functionality with session ID support
1383- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1384- Improvements to plan mode
1385
1386## 1.0.32
1387
1388- Updated loopback config for litellm
1389- Added forceLoginMethod setting to bypass login selection screen
1390
1391## 1.0.31
1392
1393- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1394
1395## 1.0.30
1396
1397- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1398- Improved file path autocomplete with filename matching
1399- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1400- Enhanced jq regex support for complex filters with pipes and select
1401
1402## 1.0.29
1403
1404- Improved CJK character support in cursor navigation and rendering
1405
1406## 1.0.28
1407
1408- Slash commands: Fix selector display during history navigation
1409- Resizes images before upload to prevent API size limit errors
1410- Added XDG_CONFIG_HOME support to configuration directory
1411- Performance optimizations for memory usage
1412- New attributes (terminal.type, language) in OpenTelemetry logging
1413
1414## 1.0.27
1415
1416- Streamable HTTP MCP servers are now supported
1417- Remote MCP servers (SSE and HTTP) now support OAuth
1418- MCP resources can now be @-mentioned
1419- /resume slash command to switch conversations within Claude Code
1420
1421## 1.0.25
1422
1423- Slash commands: moved "project" and "user" prefixes to descriptions
1424- Slash commands: improved reliability for command discovery
1425- Improved support for Ghostty
1426- Improved web search reliability
1427
1428## 1.0.24
1429
1430- Improved /mcp output
1431- Fixed a bug where settings arrays got overwritten instead of merged
1432
1433## 1.0.23
1434
1435- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1436- Released Python SDK: pip install claude-code-sdk to get started
1437
1438## 1.0.22
1439
1440- SDK: Renamed `total_cost` to `total_cost_usd`
1441
1442## 1.0.21
1443
1444- Improved editing of files with tab-based indentation
1445- Fix for tool_use without matching tool_result errors
1446- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1447
1448## 1.0.18
1449
1450- Added --add-dir CLI argument for specifying additional working directories
1451- Added streaming input support without require -p flag
1452- Improved startup performance and session storage performance
1453- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1454- Added detailed MCP server tools display (/mcp)
1455- MCP authentication and permission improvements
1456- Added auto-reconnection for MCP SSE connections on disconnect
1457- Fixed issue where pasted content was lost when dialogs appeared
1458
1459## 1.0.17
1460
1461- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1462- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1463- MCP server list UI improvements
1464- Update Claude Code process title to display "claude" instead of "node"
1465
1466## 1.0.11
1467
1468- Claude Code can now also be used with a Claude Pro subscription
1469- Added /upgrade for smoother switching to Claude Max plans
1470- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1471- Improved shell configuration error handling
1472- Improved todo list handling during compaction
1473
1474## 1.0.10
1475
1476- Added markdown table support
1477- Improved streaming performance
1478
1479## 1.0.8
1480
1481- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1482- Increased default otel interval from 1s -> 5s
1483- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1484- Fixed a regression where search tools unnecessarily asked for permissions
1485- Added support for triggering thinking non-English languages
1486- Improved compacting UI
1487
1488## 1.0.7
1489
1490- Renamed /allowed-tools -> /permissions
1491- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1492- Deprecated claude config commands in favor of editing settings.json
1493- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1494- Improved error handling for /install-github-app
1495- Bugfixes, UI polish, and tool reliability improvements
1496
1497## 1.0.6
1498
1499- Improved edit reliability for tab-indented files
1500- Respect CLAUDE_CONFIG_DIR everywhere
1501- Reduced unnecessary tool permission prompts
1502- Added support for symlinks in @file typeahead
1503- Bugfixes, UI polish, and tool reliability improvements
1504
1505## 1.0.4
1506
1507- Fixed a bug where MCP tool errors weren't being parsed correctly
1508
1509## 1.0.1
1510
1511- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1512- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1513- Updated documentation links and OAuth process descriptions
1514
1515## 1.0.0
1516
1517- Claude Code is now generally available
1518- Introducing Sonnet 4 and Opus 4 models
1519
1520## 0.2.125
1521
1522- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1523- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1524
1525## 0.2.117
1526
1527- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1528- Introduced settings.cleanupPeriodDays
1529- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1530- Introduced --debug mode
1531
1532## 0.2.108
1533
1534- You can now send messages to Claude while it works to steer Claude in real-time
1535- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1536- Fixed a bug where thinking was not working in -p mode
1537- Fixed a regression in /cost reporting
1538- Deprecated MCP wizard interface in favor of other MCP commands
1539- Lots of other bugfixes and improvements
1540
1541## 0.2.107
1542
1543- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1544
1545## 0.2.106
1546
1547- MCP SSE server configs can now specify custom headers
1548- Fixed a bug where MCP permission prompt didn't always show correctly
1549
1550## 0.2.105
1551
1552- Claude can now search the web
1553- Moved system & account status to /status
1554- Added word movement keybindings for Vim
1555- Improved latency for startup, todo tool, and file edits
1556
1557## 0.2.102
1558
1559- Improved thinking triggering reliability
1560- Improved @mention reliability for images and folders
1561- You can now paste multiple large chunks into one prompt
1562
1563## 0.2.100
1564
1565- Fixed a crash caused by a stack overflow error
1566- Made db storage optional; missing db support disables --continue and --resume
1567
1568## 0.2.98
1569
1570- Fixed an issue where auto-compact was running twice
1571
1572## 0.2.96
1573
1574- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1575
1576## 0.2.93
1577
1578- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1579- Claude now has access to a Todo list that helps it stay on track and be more organized
1580
1581## 0.2.82
1582
1583- Added support for --disallowedTools
1584- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1585
1586## 0.2.75
1587
1588- Hit Enter to queue up additional messages while Claude is working
1589- Drag in or copy/paste image files directly into the prompt
1590- @-mention files to directly add them to context
1591- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1592- Improved performance for filename auto-complete
1593
1594## 0.2.74
1595
1596- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1597- Task tool can now perform writes and run bash commands
1598
1599## 0.2.72
1600
1601- Updated spinner to indicate tokens loaded and tool usage
1602
1603## 0.2.70
1604
1605- Network commands like curl are now available for Claude to use
1606- Claude can now run multiple web queries in parallel
1607- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1608
1609## 0.2.69
1610
1611- Fixed UI glitches with improved Select component behavior
1612- Enhanced terminal output display with better text truncation logic
1613
1614## 0.2.67
1615
1616- Shared project permission rules can be saved in .claude/settings.json
1617
1618## 0.2.66
1619
1620- Print mode (-p) now supports streaming output via --output-format=stream-json
1621- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1622
1623## 0.2.63
1624
1625- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1626
1627## 0.2.61
1628
1629- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1630- Enhanced image detection for more reliable clipboard paste functionality
1631- Fixed an issue where ESC key could crash the conversation history selector
1632
1633## 0.2.59
1634
1635- Copy+paste images directly into your prompt
1636- Improved progress indicators for bash and fetch tools
1637- Bugfixes for non-interactive mode (-p)
1638
1639## 0.2.54
1640
1641- Quickly add to Memory by starting your message with '#'
1642- Press ctrl+r to see full output for long tool results
1643- Added support for MCP SSE transport
1644
1645## 0.2.53
1646
1647- New web fetch tool lets Claude view URLs that you paste in
1648- Fixed a bug with JPEG detection
1649
1650## 0.2.50
1651
1652- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1653
1654## 0.2.49
1655
1656- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1657
1658## 0.2.47
1659
1660- Press Tab to auto-complete file and folder names
1661- Press Shift + Tab to toggle auto-accept for file edits
1662- Automatic conversation compaction for infinite conversation length (toggle with /config)
1663
1664## 0.2.44
1665
1666- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1667
1668## 0.2.41
1669
1670- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1671- MCP server startup no longer blocks the app from starting up
1672
1673## 0.2.37
1674
1675- New /release-notes command lets you view release notes at any time
1676- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1677
1678## 0.2.36
1679
1680- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1681- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1682
1683## 0.2.34
1684
1685- Vim bindings for text input - enable with /vim or /config
1686
1687## 0.2.32
1688
1689- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1690- Fix for some PersistentShell issues
1691
1692## 0.2.31
1693
1694- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1695- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1696
1697## 0.2.30
1698
1699- Added ANSI color theme for better terminal compatibility
1700- Fixed issue where slash command arguments weren't being sent properly
1701- (Mac-only) API keys are now stored in macOS Keychain
1702
1703## 0.2.26
1704
1705- New /approved-tools command for managing tool permissions
1706- Word-level diff display for improved code readability
1707- Fuzzy matching for slash commands
1708
1709## 0.2.21
1710
1711- Fuzzy matching for /commands