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 6a2936a · on Jan 14, 2026 History
1,394 lines · 68.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.7
4
5- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
6- Added ability to provide feedback when accepting permission prompts
7- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
8- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
9- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
10- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
11- 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)
12- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
13- Fixed terminal title animation jitter by using fixed-width braille characters
14- Fixed plugins with git submodules not being fully initialized when installed
15- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
16- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
17- 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.
18- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
19- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
20
21## 2.1.6
22
23- Added search functionality to `/config` command for quickly filtering settings
24- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
25- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
26- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
27- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
28- Added an error display when the editor fails during Ctrl+G
29- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
30- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
31- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
32- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
33- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
34- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
35- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
36- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
37- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
38- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
39- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
40- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
41- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
42- Improved @ autocomplete with icons for different suggestion types and single-line formatting
43- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
44- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
45- Changed terminal title to "Claude Code" on startup for better window identification
46- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
47- [VSCode] Fixed usage indicator not updating after manual compact
48
49## 2.1.5
50
51- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
52
53## 2.1.4
54
55- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
56- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
57
58## 2.1.3
59
60- Merged slash commands and skills, simplifying the mental model with no change in behavior
61- Added release channel (`stable` or `latest`) toggle to `/config`
62- 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
63- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
64- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
65- Fixed mismatch between background task count in status bar and items shown in tasks dialog
66- Fixed sub-agents using the wrong model during conversation compaction
67- Fixed web search in sub-agents using incorrect model
68- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
69- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
70- Improved slash command suggestion readability by truncating long descriptions to 2 lines
71- Changed tool hook execution timeout from 60 seconds to 10 minutes
72- [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)
73
74## 2.1.2
75
76- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
77- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
78- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
79- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
80- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
81- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
82- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
83- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
84- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
85- Fixed updates incorrectly claiming another installation is in progress
86- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
87- Fixed remote session URL and teleport being broken when using `/tasks` command
88- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
89- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
90- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
91- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
92- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
93- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
94- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
95- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
96- [SDK] Changed minimum zod peer dependency to ^4.0.0
97- [VSCode] Fixed usage display not updating after manual compact
98
99## 2.1.0
100
101- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
102- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
103- Added support for `agent` field in skills to specify agent type for execution
104- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
105- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
106- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
107- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
108- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
109- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
110- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
111- Fixed Esc key with queued prompts to only move them to input without canceling the running task
112- Reduced permission prompts for complex bash commands
113- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
114- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
115- Fixed plugin path resolution for file-based marketplace sources
116- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
117- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
118- Fixed Claude in Chrome support for WSL environments
119- Fixed Windows native installer silently failing when executable creation fails
120- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
121- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
122- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
123- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
124- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
125- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
126- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
127- Added hooks support for skill and slash command frontmatter
128- 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
129- Added `/plan` command shortcut to enable plan mode directly from the prompt
130- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
131- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
132- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
133- Added support for `once: true` config for hooks
134- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
135- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
136- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
137- Added left/right arrow key navigation for cycling through tabs in dialogs
138- Added real-time thinking block display in Ctrl+O transcript mode
139- Added filepath to full output in background bash task details dialog
140- Added Skills as a separate category in the context visualization
141- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
142- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
143- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
144- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
145- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
146- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
147- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
148- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
149- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
150- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
151- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
152- Fixed commands with `$()` command substitution failing with parse errors
153- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
154- 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)
155- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
156- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
157- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
158- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
159- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
160- Fixed subagents sometimes not inheriting the parent's model by default
161- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
162- Fixed duplicate Bash commands appearing in permission request option labels
163- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
164- Fixed background task completion notifications to appear proactively with bullet point
165- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
166- Fixed cursor disappearing after dismissing permission dialogs
167- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
168- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
169- Fixed images being silently dropped when queueing messages while backgrounding a task
170- Fixed large pasted images failing with "Image was too large" error
171- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
172- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
173- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
174- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
175- Fixed spinner token counter to properly accumulate tokens from subagents during execution
176- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
177- Fixed race condition where LSP tool could return "no server available" during startup
178- Fixed feedback submission hanging indefinitely when network requests timeout
179- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
180- Fixed hook success message showing trailing colon when hook has no output
181- Multiple optimizations to improve startup performance
182- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
183- Improved performance when reading Jupyter notebooks with many cells
184- Improved reliability for piped input like `cat refactor.md | claude`
185- Improved reliability for AskQuestion tool
186- Improved sed in-place edit commands to render as file edits with diff preview
187- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
188- Improved compaction reliability
189- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
190- Improved skills to show progress while executing, displaying tool uses as they happen
191- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
192- Improved skill suggestions to prioritize recently and frequently used skills
193- Improved spinner feedback when waiting for the first response token
194- Improved token count display in spinner to include tokens from background agents
195- Improved incremental output for async agents to give the main thread more control and visibility
196- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
197- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
198- Improved macOS screenshot paste reliability with TIFF format support
199- Improved `/stats` output
200- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
201- Changed "Interrupted" message color from red to grey for a less alarming appearance
202- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
203- Removed underline styling from image reference links
204- [SDK] Changed minimum zod peer dependency to ^4.0.0
205- [VSCode] Added currently selected model name to the context menu
206- [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")
207- [VSCode] Fixed paragraph breaks not rendering in markdown content
208- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
209- [Windows] Fixed issue with improper rendering
210
211## 2.0.76
212
213- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
214
215## 2.0.75
216
217- Minor bugfixes
218
219## 2.0.74
220
221- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
222- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
223- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
224- Added syntax highlighting info to theme picker
225- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
226- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
227- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
228- Fixed a potential crash when syntax highlighting isn't initialized correctly
229- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
230- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
231- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
232- [Windows] Fixed issue with improper rendering
233- [VSCode] Added gift tag pictogram for year-end promotion message
234
235## 2.0.73
236
237- Added clickable `[Image #N]` links that open attached images in the default viewer
238- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
239- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
240- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
241- Fixed slow input history cycling and race condition that could overwrite text after message submission
242- Improved `/theme` command to open theme picker directly
243- Improved theme picker UI
244- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
245- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
246
247## 2.0.72
248
249- 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
250- Reduced terminal flickering
251- Added scannable QR code to mobile app tip for quick app downloads
252- Added loading indicator when resuming conversations for better feedback
253- Fixed `/context` command not respecting custom system prompts in non-interactive mode
254- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
255- Improved @ mention file suggestion speed (~3x faster in git repositories)
256- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
257- Improved settings validation errors to be more prominent
258- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
259
260## 2.0.71
261
262- Added /config toggle to enable/disable prompt suggestions
263- Added `/settings` as an alias for the `/config` command
264- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
265- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
266- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
267- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
268- New syntax highlighting engine for native build
269
270## 2.0.70
271
272- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
273- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
274- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
275- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
276- Fixed input being cleared when processing queued commands while the user was typing
277- Fixed prompt suggestions replacing typed input when pressing Tab
278- Fixed diff view not updating when terminal is resized
279- Improved memory usage by 3x for large conversations
280- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
281- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
282- Fix thinking mode toggle in /config not persisting correctly
283- Improve UI for file creation permission dialog
284
285## 2.0.69
286
287- Minor bugfixes
288
289## 2.0.68
290
291- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
292- Fixed a bug where disallowed MCP tools were visible to the model
293- Fixed an issue where steering messages could be lost while a subagent is working
294- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
295- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
296- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
297
298## 2.0.67
299
300- Thinking mode is now enabled by default for Opus 4.5
301- Thinking mode configuration has moved to /config
302- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
303- Show reason why autoupdater is disabled in `/doctor`
304- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
305- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
306- Fixed scroll position resetting after deleting a permission rule in `/permissions`
307- 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
308- Fixed `claude install --force` not bypassing stale lock files
309- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
310- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
311
312## 2.0.65
313
314- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
315- Added context window information to status line input
316- Added `fileSuggestion` setting for custom `@` file search commands
317- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
318- Fixed prompt not being saved to history when aborting a query with Escape
319- Fixed Read tool image handling to identify format from bytes instead of file extension
320
321## 2.0.64
322
323- Made auto-compacting instant
324- Agents and bash commands can run asynchronously and send messages to wake up the main agent
325- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
326- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
327- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
328- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
329- Fixed auto-loading .env when using native installer
330- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
331- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
332- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
333- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
334- Bedrock: Improve efficiency of token counting
335- Bedrock: Add support for `aws login` AWS Management Console credentials
336- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
337
338## 2.0.62
339
340- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
341- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
342- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
343- Fixed slash command selection not working when multiple commands share the same name
344- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
345- Fixed running versions getting removed because lock file incorrectly going stale
346- Fixed IDE diff tab not closing when rejecting file changes
347
348## 2.0.61
349
350- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
351
352## 2.0.60
353
354- Added background agent support. Agents run in the background while you work
355- Added --disable-slash-commands CLI flag to disable all slash commands
356- Added model name to "Co-Authored-By" commit messages
357- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
358- Updated Fetch to skip summarization for pre-approved websites
359- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
360
361## 2.0.59
362
363- Added --agent CLI flag to override the agent setting for the current session
364- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
365- VS Code: Fixed .claude.json config file being read from incorrect location
366
367## 2.0.58
368
369- Pro users now have access to Opus 4.5 as part of their subscription!
370- Fixed timer duration showing "11m 60s" instead of "12m 0s"
371- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
372
373## 2.0.57
374
375- Added feedback input when rejecting plans, allowing users to tell Claude what to change
376- VSCode: Added streaming message support for real-time response display
377
378## 2.0.56
379
380- Added setting to enable/disable terminal progress bar (OSC 9;4)
381- 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.
382
383## 2.0.55
384
385- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
386- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
387- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
388- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
389
390## 2.0.54
391
392- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
393- Fix issue with excessive iTerm notifications
394
395## 2.0.52
396
397- Fixed duplicate message display when starting Claude with a command line argument
398- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
399- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
400- Permit some uses of `$!` in bash commands
401
402## 2.0.51
403
404- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
405- Introducing Claude Code for Desktop: https://claude.com/download
406- 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
407- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
408- Plan Mode now builds more precise plans and executes more thoroughly
409- Usage limit notifications now easier to understand
410- Switched `/usage` back to "% used"
411- Fixed handling of thinking errors
412- Fixed performance regression
413
414## 2.0.50
415
416- Fixed bug preventing calling MCP tools that have nested references in their input schemas
417- Silenced a noisy but harmless error during upgrades
418- Improved ultrathink text display
419- Improved clarity of 5-hour session limit warning message
420
421## 2.0.49
422
423- Added readline-style ctrl-y for pasting deleted text
424- Improved clarity of usage limit warning message
425- Fixed handling of subagent permissions
426
427## 2.0.47
428
429- Improved error messages and validation for `claude --teleport`
430- Improved error handling in `/usage`
431- Fixed race condition with history entry not getting logged at exit
432- Fixed Vertex AI configuration not being applied from `settings.json`
433
434## 2.0.46
435
436- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
437
438## 2.0.45
439
440- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
441- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
442- Send background tasks to Claude Code on the web by starting a message with `&`
443
444## 2.0.43
445
446- Added `permissionMode` field for custom agents
447- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
448- Added skills frontmatter field to declare skills to auto-load for subagents
449- Added the `SubagentStart` hook event
450- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
451- Fixed duplicate rendering of some messages in the UI
452- Fixed some visual flickers
453- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
454
455## 2.0.42
456
457- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
458
459## 2.0.41
460
461- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
462- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
463- Fixed incorrect labeling of user settings vs project settings in command descriptions
464- Fixed crash when plugin command hooks timeout during execution
465- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
466- Fixed broken security documentation links in trust dialogs and onboarding
467- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
468- ctrl-r history search landing on a slash command no longer cancels the search
469- SDK: Support custom timeouts for hooks
470- Allow more safe git commands to run without approval
471- Plugins: Added support for sharing and installing output styles
472- Teleporting a session from web will automatically set the upstream branch
473
474## 2.0.37
475
476- Fixed how idleness is computed for notifications
477- Hooks: Added matcher values for Notification hook events
478- Output Styles: Added `keep-coding-instructions` option to frontmatter
479
480## 2.0.36
481
482- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
483- Fixed queued messages being incorrectly executed as bash commands
484- Fixed input being lost when typing while a queued message is processed
485
486## 2.0.35
487
488- Improve fuzzy search results when searching commands
489- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
490- 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
491- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
492- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
493
494## 2.0.34
495
496- VSCode Extension: Added setting to configure the initial permission mode for new conversations
497- Improved file path suggestion performance with native Rust-based fuzzy finder
498- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
499- Fixed memory crash when reading or writing large files (especially base64-encoded images)
500
501## 2.0.33
502
503- Native binary installs now launch quicker.
504- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
505- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
506
507## 2.0.32
508
509- Un-deprecate output styles based on community feedback
510- Added `companyAnnouncements` setting for displaying announcements on startup
511- Fixed hook progress messages not updating correctly during PostToolUse hook execution
512
513## 2.0.31
514
515- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
516- Vertex: add support for Web Search on supported models
517- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
518- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
519- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
520- Fixed plugin uninstall not removing plugins
521
522## 2.0.30
523
524- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
525- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
526- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
527- Added prompt-based stop hooks
528- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
529- Enabled SSE MCP servers on native build
530- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
531- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
532- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
533- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
534- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
535- Fixed bug that saved session permissions to local settings
536- Fixed MCP tools not being available to sub-agents
537- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
538- Fixed delay when navigating through typeahead suggestions with arrow keys
539- VSCode: Restored selection indicator in input footer showing current file or code selection status
540
541## 2.0.28
542
543- Plan mode: introduced new Plan subagent
544- Subagents: claude can now choose to resume subagents
545- Subagents: claude can dynamically choose the model used by its subagents
546- SDK: added --max-budget-usd flag
547- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
548- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
549- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
550- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
551- Various other bug fixes
552
553## 2.0.27
554
555- New UI for permission prompts
556- Added current branch filtering and search to session resume screen for easier navigation
557- Fixed directory @-mention causing "No assistant message found" error
558- VSCode Extension: Add config setting to include .gitignored files in file searches
559- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
560
561## 2.0.25
562
563- 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
564
565## 2.0.24
566
567- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
568- Claude Code Web: Support for Web -> CLI teleport
569- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
570- Bedrock: Display awsAuthRefresh output when auth is required
571
572## 2.0.22
573
574- Fixed content layout shift when scrolling through slash commands
575- IDE: Add toggle to enable/disable thinking.
576- Fix bug causing duplicate permission prompts with parallel tool calls
577- Add support for enterprise managed MCP allowlist and denylist
578
579## 2.0.21
580
581- Support MCP `structuredContent` field in tool responses
582- Added an interactive question tool
583- Claude will now ask you questions more often in plan mode
584- Added Haiku 4.5 as a model option for Pro users
585- Fixed an issue where queued commands don't have access to previous messages' output
586
587## 2.0.20
588
589- Added support for Claude Skills
590
591## 2.0.19
592
593- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
594- Fixed a bug where Haiku was unnecessarily called in print mode
595
596## 2.0.17
597
598- Added Haiku 4.5 to model selector!
599- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
600- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
601- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
602- OTEL: support HTTP_PROXY and HTTPS_PROXY
603- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
604
605## 2.0.15
606
607- Fixed bug with resuming where previously created files needed to be read again before writing
608- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
609
610## 2.0.14
611
612- Fix @-mentioning MCP servers to toggle them on/off
613- Improve permission checks for bash with inline env vars
614- Fix ultrathink + thinking toggle
615- Reduce unnecessary logins
616- Document --system-prompt
617- Several improvements to rendering
618- Plugins UI polish
619
620## 2.0.13
621
622- Fixed `/plugin` not working on native build
623
624## 2.0.12
625
626- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
627- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
628- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
629- `/plugin validate` command for validating plugin structure and configuration
630- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
631- Plugin documentation available at https://code.claude.com/docs/en/plugins
632- Comprehensive error messages and diagnostics via `/doctor` command
633- Avoid flickering in `/model` selector
634- Improvements to `/help`
635- Avoid mentioning hooks in `/resume` summaries
636- Changes to the "verbose" setting in `/config` now persist across sessions
637
638## 2.0.11
639
640- Reduced system prompt size by 1.4k tokens
641- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
642- Fixed Opus fallback rate limit errors appearing incorrectly
643- Fixed /add-dir command selecting wrong default tab
644
645## 2.0.10
646
647- Rewrote terminal renderer for buttery smooth UI
648- Enable/disable MCP servers by @mentioning, or in /mcp
649- Added tab completion for shell commands in bash mode
650- PreToolUse hooks can now modify tool inputs
651- Press Ctrl-G to edit your prompt in your system's configured text editor
652- Fixes for bash permission checks with environment variables in the command
653
654## 2.0.9
655
656- Fix regression where bash backgrounding stopped working
657
658## 2.0.8
659
660- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
661- IDE: Add drag-and-drop support for files and folders in chat
662- /context: Fix counting for thinking blocks
663- Improve message rendering for users with light themes on dark terminals
664- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
665
666## 2.0.5
667
668- IDE: Fix IME unintended message submission with Enter and Tab
669- IDE: Add "Open in Terminal" link in login screen
670- Fix unhandled OAuth expiration 401 API errors
671- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
672
673## 2.0.1
674
675- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
676- Various bug fixes and presentation improvements
677
678## 2.0.0
679
680- New native VS Code extension
681- Fresh coat of paint throughout the whole app
682- /rewind a conversation to undo code changes
683- /usage command to see plan limits
684- Tab to toggle thinking (sticky across sessions)
685- Ctrl-R to search history
686- Unshipped claude config command
687- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
688- SDK: The Claude Code SDK is now the Claude Agent SDK
689- Add subagents dynamically with `--agents` flag
690
691## 1.0.126
692
693- Enable /context command for Bedrock and Vertex
694- Add mTLS support for HTTP-based OpenTelemetry exporters
695
696## 1.0.124
697
698- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
699- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
700- No longer inform Claude of the list of allowed tools when permission is denied
701- Fixed security vulnerability in Bash tool permission checks
702- Improved VSCode extension performance for large files
703
704## 1.0.123
705
706- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
707- Fixed thinking mode triggering on negation phrases like "don't think"
708- Fixed rendering performance degradation during token streaming
709- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
710- Enhanced BashTool environment snapshot logging
711- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
712- Migrated --debug logging to a file, to enable easy tailing & filtering
713
714## 1.0.120
715
716- Fix input lag during typing, especially noticeable with large prompts
717- Improved VSCode extension command registry and sessions dialog user experience
718- Enhanced sessions dialog responsiveness and visual feedback
719- Fixed IDE compatibility issue by removing worktree support check
720- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
721
722## 1.0.119
723
724- Fix Windows issue where process visually freezes on entering interactive mode
725- Support dynamic headers for MCP servers via headersHelper configuration
726- Fix thinking mode not working in headless sessions
727- Fix slash commands now properly update allowed tools instead of replacing them
728
729## 1.0.117
730
731- Add Ctrl-R history search to recall previous commands like bash/zsh
732- Fix input lag while typing, especially on Windows
733- Add sed command to auto-allowed commands in acceptEdits mode
734- Fix Windows PATH comparison to be case-insensitive for drive letters
735- Add permissions management hint to /add-dir output
736
737## 1.0.115
738
739- Improve thinking mode display with enhanced visual effects
740- Type /t to temporarily disable thinking mode in your prompt
741- Improve path validation for glob and grep tools
742- Show condensed output for post-tool hooks to reduce visual clutter
743- Fix visual feedback when loading state completes
744- Improve UI consistency for permission request dialogs
745
746## 1.0.113
747
748- Deprecated piped input in interactive mode
749- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
750
751## 1.0.112
752
753- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
754- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
755- Hooks: Added systemMessage support for SessionEnd hooks
756- Added `spinnerTipsEnabled` setting to disable spinner tips
757- IDE: Various improvements and bug fixes
758
759## 1.0.111
760
761- /model now validates provided model names
762- Fixed Bash tool crashes caused by malformed shell syntax parsing
763
764## 1.0.110
765
766- /terminal-setup command now supports WezTerm
767- MCP: OAuth tokens now proactively refresh before expiration
768- Fixed reliability issues with background Bash processes
769
770## 1.0.109
771
772- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
773
774## 1.0.106
775
776- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
777
778## 1.0.97
779
780- Settings: /doctor now validates permission rule syntax and suggests corrections
781
782## 1.0.94
783
784- Vertex: add support for global endpoints for supported models
785- /memory command now allows direct editing of all imported memory files
786- SDK: Add custom tools as callbacks
787- Added /todos command to list current todo items
788
789## 1.0.93
790
791- Windows: Add alt + v shortcut for pasting images from clipboard
792- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
793
794## 1.0.90
795
796- Settings file changes take effect immediately - no restart required
797
798## 1.0.88
799
800- Fixed issue causing "OAuth authentication is currently not supported"
801- Status line input now includes `exceeds_200k_tokens`
802- Fixed incorrect usage tracking in /cost.
803- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
804- Bedrock: Updated default Sonnet model to Sonnet 4
805
806## 1.0.86
807
808- Added /context to help users self-serve debug context issues
809- SDK: Added UUID support for all SDK messages
810- SDK: Added `--replay-user-messages` to replay user messages back to stdout
811
812## 1.0.85
813
814- Status line input now includes session cost info
815- Hooks: Introduced SessionEnd hook
816
817## 1.0.84
818
819- Fix tool_use/tool_result id mismatch error when network is unstable
820- Fix Claude sometimes ignoring real-time steering when wrapping up a task
821- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
822- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
823
824## 1.0.83
825
826- @-mention: Support files with spaces in path
827- New shimmering spinner
828
829## 1.0.82
830
831- SDK: Add request cancellation support
832- SDK: New additionalDirectories option to search custom paths, improved slash command processing
833- Settings: Validation prevents invalid fields in .claude/settings.json files
834- MCP: Improve tool name consistency
835- Bash: Fix crash when Claude tries to automatically read large files
836
837## 1.0.81
838
839- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
840- Agents: Fix custom agent loading when agent files are unparsable
841
842## 1.0.80
843
844- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
845
846## 1.0.77
847
848- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
849- SDK: Add session support and permission denial tracking
850- Fix token limit errors in conversation summarization
851- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
852
853## 1.0.73
854
855- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
856- MCP: Press Esc to cancel OAuth authentication flows
857- Bash: Improved command validation and reduced false security warnings
858- UI: Enhanced spinner animations and status line visual hierarchy
859- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
860
861## 1.0.72
862
863- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
864
865## 1.0.71
866
867- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
868- Customizable status line: add your terminal prompt to Claude Code with /statusline
869
870## 1.0.70
871
872- Performance: Optimized message rendering for better performance with large contexts
873- Windows: Fixed native file search, ripgrep, and subagent functionality
874- Added support for @-mentions in slash command arguments
875
876## 1.0.69
877
878- Upgraded Opus to version 4.1
879
880## 1.0.68
881
882- Fix incorrect model names being used for certain commands like `/pr-comments`
883- 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.
884- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
885- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
886- SDK: Added canUseTool callback support for tool confirmation
887- Added `disableAllHooks` setting
888- Improved file suggestions performance in large repos
889
890## 1.0.65
891
892- IDE: Fixed connection stability issues and error handling for diagnostics
893- Windows: Fixed shell environment setup for users without .bashrc files
894
895## 1.0.64
896
897- Agents: Added model customization support - you can now specify which model an agent should use
898- Agents: Fixed unintended access to the recursive agent tool
899- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
900- SDK: Fixed user input tracking across multi-turn conversations
901- Added hidden files to file search and @-mention suggestions
902
903## 1.0.63
904
905- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
906
907## 1.0.62
908
909- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
910- Hooks: Added SessionStart hook for new session initialization
911- /add-dir command now supports typeahead for directory paths
912- Improved network connectivity check reliability
913
914## 1.0.61
915
916- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
917- Settings: Added `--settings` flag to load settings from a JSON file
918- Settings: Fixed resolution of settings files paths that are symlinks
919- OTEL: Fixed reporting of wrong organization after authentication changes
920- Slash commands: Fixed permissions checking for allowed-tools with Bash
921- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
922- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
923- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
924
925## 1.0.60
926
927- You can now create custom subagents for specialized tasks! Run /agents to get started
928
929## 1.0.59
930
931- SDK: Added tool confirmation support with canUseTool callback
932- SDK: Allow specifying env for spawned process
933- Hooks: Exposed PermissionDecision to hooks (including "ask")
934- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
935- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
936
937## 1.0.58
938
939- Added support for reading PDFs
940- MCP: Improved server health status display in 'claude mcp list'
941- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
942
943## 1.0.57
944
945- Added support for specifying a model in slash commands
946- Improved permission messages to help Claude understand allowed tools
947- Fix: Remove trailing newlines from bash output in terminal wrapping
948
949## 1.0.56
950
951- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
952- Fixes for WSL IDE detection
953- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
954
955## 1.0.55
956
957- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
958- Windows: fixed Ctrl+Z crash
959- SDK: Added ability to capture error logging
960- Add --system-prompt-file option to override system prompt in print mode
961
962## 1.0.54
963
964- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
965- Custom slash commands: Added argument-hint to frontmatter
966- Windows: OAuth uses port 45454 and properly constructs browser URL
967- Windows: mode switching now uses alt + m, and plan mode renders properly
968- Shell: Switch to in-memory shell snapshot to fix file-related errors
969
970## 1.0.53
971
972- Updated @-mention file truncation from 100 lines to 2000 lines
973- 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).
974
975## 1.0.52
976
977- Added support for MCP server instructions
978
979## 1.0.51
980
981- Added support for native Windows (requires Git for Windows)
982- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
983- Settings: /doctor can now help you identify and fix invalid setting files
984- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
985- Increased auto-compact warning threshold from 60% to 80%
986- Fixed an issue with handling user directories with spaces for shell snapshots
987- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
988- Custom slash commands: Fixed user-level commands in subdirectories
989- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
990
991## 1.0.48
992
993- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
994- Added progress messages to Bash tool based on the last 5 lines of command output
995- Added expanding variables support for MCP server configuration
996- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
997- Improved IDE extension path handling when Claude Code runs in WSL
998- Hooks: Added a PreCompact hook
999- Vim mode: Added c, f/F, t/T
1000
1001## 1.0.45
1002
1003- Redesigned Search (Grep) tool with new tool input parameters and features
1004- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1005- Fixed config file corruption issue by enforcing atomic writes
1006- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1007- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1008- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1009
1010## 1.0.44
1011
1012- New /export command lets you quickly export a conversation for sharing
1013- MCP: resource_link tool results are now supported
1014- MCP: tool annotations and tool titles now display in /mcp view
1015- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1016
1017## 1.0.43
1018
1019- Fixed a bug where the theme selector was saving excessively
1020- Hooks: Added EPIPE system error handling
1021
1022## 1.0.42
1023
1024- Added tilde (`~`) expansion support to `/add-dir` command
1025
1026## 1.0.41
1027
1028- Hooks: Split Stop hook triggering into Stop and SubagentStop
1029- Hooks: Enabled optional timeout configuration for each command
1030- Hooks: Added "hook_event_name" to hook input
1031- Fixed a bug where MCP tools would display twice in tool list
1032- New tool parameters JSON for Bash tool in `tool_decision` event
1033
1034## 1.0.40
1035
1036- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1037
1038## 1.0.39
1039
1040- New Active Time metric in OpenTelemetry logging
1041
1042## 1.0.38
1043
1044- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1045
1046## 1.0.37
1047
1048- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1049
1050## 1.0.36
1051
1052- Web search now takes today's date into context
1053- Fixed a bug where stdio MCP servers were not terminating properly on exit
1054
1055## 1.0.35
1056
1057- Added support for MCP OAuth Authorization Server discovery
1058
1059## 1.0.34
1060
1061- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1062
1063## 1.0.33
1064
1065- Improved logging functionality with session ID support
1066- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1067- Improvements to plan mode
1068
1069## 1.0.32
1070
1071- Updated loopback config for litellm
1072- Added forceLoginMethod setting to bypass login selection screen
1073
1074## 1.0.31
1075
1076- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1077
1078## 1.0.30
1079
1080- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1081- Improved file path autocomplete with filename matching
1082- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1083- Enhanced jq regex support for complex filters with pipes and select
1084
1085## 1.0.29
1086
1087- Improved CJK character support in cursor navigation and rendering
1088
1089## 1.0.28
1090
1091- Slash commands: Fix selector display during history navigation
1092- Resizes images before upload to prevent API size limit errors
1093- Added XDG_CONFIG_HOME support to configuration directory
1094- Performance optimizations for memory usage
1095- New attributes (terminal.type, language) in OpenTelemetry logging
1096
1097## 1.0.27
1098
1099- Streamable HTTP MCP servers are now supported
1100- Remote MCP servers (SSE and HTTP) now support OAuth
1101- MCP resources can now be @-mentioned
1102- /resume slash command to switch conversations within Claude Code
1103
1104## 1.0.25
1105
1106- Slash commands: moved "project" and "user" prefixes to descriptions
1107- Slash commands: improved reliability for command discovery
1108- Improved support for Ghostty
1109- Improved web search reliability
1110
1111## 1.0.24
1112
1113- Improved /mcp output
1114- Fixed a bug where settings arrays got overwritten instead of merged
1115
1116## 1.0.23
1117
1118- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1119- Released Python SDK: pip install claude-code-sdk to get started
1120
1121## 1.0.22
1122
1123- SDK: Renamed `total_cost` to `total_cost_usd`
1124
1125## 1.0.21
1126
1127- Improved editing of files with tab-based indentation
1128- Fix for tool_use without matching tool_result errors
1129- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1130
1131## 1.0.18
1132
1133- Added --add-dir CLI argument for specifying additional working directories
1134- Added streaming input support without require -p flag
1135- Improved startup performance and session storage performance
1136- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1137- Added detailed MCP server tools display (/mcp)
1138- MCP authentication and permission improvements
1139- Added auto-reconnection for MCP SSE connections on disconnect
1140- Fixed issue where pasted content was lost when dialogs appeared
1141
1142## 1.0.17
1143
1144- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1145- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1146- MCP server list UI improvements
1147- Update Claude Code process title to display "claude" instead of "node"
1148
1149## 1.0.11
1150
1151- Claude Code can now also be used with a Claude Pro subscription
1152- Added /upgrade for smoother switching to Claude Max plans
1153- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1154- Improved shell configuration error handling
1155- Improved todo list handling during compaction
1156
1157## 1.0.10
1158
1159- Added markdown table support
1160- Improved streaming performance
1161
1162## 1.0.8
1163
1164- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1165- Increased default otel interval from 1s -> 5s
1166- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1167- Fixed a regression where search tools unnecessarily asked for permissions
1168- Added support for triggering thinking non-English languages
1169- Improved compacting UI
1170
1171## 1.0.7
1172
1173- Renamed /allowed-tools -> /permissions
1174- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1175- Deprecated claude config commands in favor of editing settings.json
1176- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1177- Improved error handling for /install-github-app
1178- Bugfixes, UI polish, and tool reliability improvements
1179
1180## 1.0.6
1181
1182- Improved edit reliability for tab-indented files
1183- Respect CLAUDE_CONFIG_DIR everywhere
1184- Reduced unnecessary tool permission prompts
1185- Added support for symlinks in @file typeahead
1186- Bugfixes, UI polish, and tool reliability improvements
1187
1188## 1.0.4
1189
1190- Fixed a bug where MCP tool errors weren't being parsed correctly
1191
1192## 1.0.1
1193
1194- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1195- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1196- Updated documentation links and OAuth process descriptions
1197
1198## 1.0.0
1199
1200- Claude Code is now generally available
1201- Introducing Sonnet 4 and Opus 4 models
1202
1203## 0.2.125
1204
1205- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1206- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1207
1208## 0.2.117
1209
1210- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1211- Introduced settings.cleanupPeriodDays
1212- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1213- Introduced --debug mode
1214
1215## 0.2.108
1216
1217- You can now send messages to Claude while it works to steer Claude in real-time
1218- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1219- Fixed a bug where thinking was not working in -p mode
1220- Fixed a regression in /cost reporting
1221- Deprecated MCP wizard interface in favor of other MCP commands
1222- Lots of other bugfixes and improvements
1223
1224## 0.2.107
1225
1226- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1227
1228## 0.2.106
1229
1230- MCP SSE server configs can now specify custom headers
1231- Fixed a bug where MCP permission prompt didn't always show correctly
1232
1233## 0.2.105
1234
1235- Claude can now search the web
1236- Moved system & account status to /status
1237- Added word movement keybindings for Vim
1238- Improved latency for startup, todo tool, and file edits
1239
1240## 0.2.102
1241
1242- Improved thinking triggering reliability
1243- Improved @mention reliability for images and folders
1244- You can now paste multiple large chunks into one prompt
1245
1246## 0.2.100
1247
1248- Fixed a crash caused by a stack overflow error
1249- Made db storage optional; missing db support disables --continue and --resume
1250
1251## 0.2.98
1252
1253- Fixed an issue where auto-compact was running twice
1254
1255## 0.2.96
1256
1257- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1258
1259## 0.2.93
1260
1261- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1262- Claude now has access to a Todo list that helps it stay on track and be more organized
1263
1264## 0.2.82
1265
1266- Added support for --disallowedTools
1267- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1268
1269## 0.2.75
1270
1271- Hit Enter to queue up additional messages while Claude is working
1272- Drag in or copy/paste image files directly into the prompt
1273- @-mention files to directly add them to context
1274- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1275- Improved performance for filename auto-complete
1276
1277## 0.2.74
1278
1279- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1280- Task tool can now perform writes and run bash commands
1281
1282## 0.2.72
1283
1284- Updated spinner to indicate tokens loaded and tool usage
1285
1286## 0.2.70
1287
1288- Network commands like curl are now available for Claude to use
1289- Claude can now run multiple web queries in parallel
1290- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1291
1292## 0.2.69
1293
1294- Fixed UI glitches with improved Select component behavior
1295- Enhanced terminal output display with better text truncation logic
1296
1297## 0.2.67
1298
1299- Shared project permission rules can be saved in .claude/settings.json
1300
1301## 0.2.66
1302
1303- Print mode (-p) now supports streaming output via --output-format=stream-json
1304- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1305
1306## 0.2.63
1307
1308- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1309
1310## 0.2.61
1311
1312- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1313- Enhanced image detection for more reliable clipboard paste functionality
1314- Fixed an issue where ESC key could crash the conversation history selector
1315
1316## 0.2.59
1317
1318- Copy+paste images directly into your prompt
1319- Improved progress indicators for bash and fetch tools
1320- Bugfixes for non-interactive mode (-p)
1321
1322## 0.2.54
1323
1324- Quickly add to Memory by starting your message with '#'
1325- Press ctrl+r to see full output for long tool results
1326- Added support for MCP SSE transport
1327
1328## 0.2.53
1329
1330- New web fetch tool lets Claude view URLs that you paste in
1331- Fixed a bug with JPEG detection
1332
1333## 0.2.50
1334
1335- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1336
1337## 0.2.49
1338
1339- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1340
1341## 0.2.47
1342
1343- Press Tab to auto-complete file and folder names
1344- Press Shift + Tab to toggle auto-accept for file edits
1345- Automatic conversation compaction for infinite conversation length (toggle with /config)
1346
1347## 0.2.44
1348
1349- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1350
1351## 0.2.41
1352
1353- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1354- MCP server startup no longer blocks the app from starting up
1355
1356## 0.2.37
1357
1358- New /release-notes command lets you view release notes at any time
1359- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1360
1361## 0.2.36
1362
1363- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1364- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1365
1366## 0.2.34
1367
1368- Vim bindings for text input - enable with /vim or /config
1369
1370## 0.2.32
1371
1372- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1373- Fix for some PersistentShell issues
1374
1375## 0.2.31
1376
1377- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1378- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1379
1380## 0.2.30
1381
1382- Added ANSI color theme for better terminal compatibility
1383- Fixed issue where slash command arguments weren't being sent properly
1384- (Mac-only) API keys are now stored in macOS Keychain
1385
1386## 0.2.26
1387
1388- New /approved-tools command for managing tool permissions
1389- Word-level diff display for improved code readability
1390- Fuzzy matching for slash commands
1391
1392## 0.2.21
1393
1394- Fuzzy matching for /commands