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 6aecb15 · on Feb 24, 2026 History
1,879 lines · 109.2 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.51
4
5- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
6- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
7- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
8- BashTool now skips login shell (`-l` flag) by default when a shell snapshot is available, improving command execution performance. Previously this required setting `CLAUDE_BASH_NO_LOGIN=true`.
9- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
10- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
11- Fixed a security issue where HTTP hooks could interpolate arbitrary environment variables from header values. Env var interpolation now requires an explicit `allowedEnvVars` list in the hook configuration.
12- Fixed a bug where duplicate `control_response` messages (e.g. from WebSocket reconnects) could cause API 400 errors by pushing duplicate assistant messages into the conversation.
13- Added `CLAUDE_CODE_ACCOUNT_UUID`, `CLAUDE_CODE_USER_EMAIL`, and `CLAUDE_CODE_ORGANIZATION_UUID` environment variables for SDK callers to provide account info synchronously, eliminating a race condition where early telemetry events lacked account metadata.
14- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
15- HTTP hooks are now routed through the sandbox network proxy when sandboxing is enabled, enforcing the domain allowlist. HTTP hooks are not supported for SessionStart/Setup events.
16- The `/model` picker now shows human-readable labels (e.g., "Sonnet 4.5") instead of raw model IDs for pinned model versions, with an upgrade hint when a newer version is available.
17
18## 2.1.50
19
20- Added support for `startupTimeout` configuration for LSP servers
21- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
22- Fixed a bug where resumed sessions could be invisible when the working directory involved symlinks, because the session storage path was resolved at different times during startup. Also fixed session data loss on SSH disconnect by flushing session data before hooks and analytics in the graceful shutdown sequence.
23- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
24- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
25- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
26- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
27- Fixed memory leak where completed task state objects were never removed from AppState
28- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
29- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
30- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
31- Improved memory usage during long sessions by clearing internal caches after compaction
32- Added `claude agents` CLI command to list all configured agents
33- Improved memory usage during long sessions by clearing large tool results after they have been processed
34- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
35- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
36- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
37- Fixed prompt suggestion cache regression that reduced cache hit rates
38- Fixed unbounded memory growth in long sessions by capping file history snapshots
39- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
40- Opus 4.6 (fast mode) now includes the full 1M context window
41- VSCode: Added `/extra-usage` command support in VS Code sessions
42- Fixed memory leak where TaskOutput retained recent lines after cleanup
43- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
44- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
45
46## 2.1.49
47
48- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
49- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
50- Subagents support `isolation: "worktree"` for working in a temporary git worktree
51- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
52- Agent definitions support `background: true` to always run as a background task
53- Plugins can ship `settings.json` for default configuration
54- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
55- Fixed Ctrl+C and ESC being silently ignored when background agents are running and the main thread is idle. Pressing twice within 3 seconds now kills all background agents.
56- Fixed prompt suggestion cache regression that reduced cache hit rates.
57- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
58- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
59- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
60- Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in /model.
61- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
62- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
63- Fixed potential rendering issues caused by stale yoga layout references
64- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
65- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
66- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
67- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
68- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
69- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
70- Improved startup performance by reducing HTTP calls for analytics token counting
71- Improved startup performance by batching MCP tool token counting into a single API call
72- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
73- Fixed `--resume` session picker showing raw XML tags for sessions that start with commands like `/clear`. Now correctly falls through to the session ID fallback.
74- Improved permission prompts for path safety and working directory blocks to show the reason for the restriction instead of a bare prompt with no context
75
76## 2.1.47
77
78- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
79- Fixed Windows terminal rendering bugs caused by `os.EOL` (`\r\n`) in display code — line counts now show correct values instead of always showing 1 on Windows.
80- Improved VS Code plan preview: auto-updates as Claude iterates, enables commenting only when the plan is ready for review, and keeps the preview open when rejecting so Claude can revise.
81- Fixed a bug where bold and colored text in markdown output could shift to the wrong characters on Windows due to `\r\n` line endings.
82- Fixed compaction failing when conversation contains many PDF documents by stripping document blocks alongside images before sending to the compaction API (anthropics/claude-code#26188)
83- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
84- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
85- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
86- Improved performance of `@` file mentions - file suggestions now appear faster by pre-warming the index on startup and using session-based caching with background refresh.
87- Improved memory usage by trimming agent task message history after tasks complete
88- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
89- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
90- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
91- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
92- Search patterns in collapsed tool results are now displayed in quotes for clarity
93- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
94- Use `ctrl+f` to kill all background agents instead of double-pressing ESC. Background agents now continue running when you press ESC to cancel the main thread, giving you more control over agent lifecycle.
95- Fixed API 400 errors ("thinking blocks cannot be modified") that occurred in sessions with concurrent agents, caused by interleaved streaming content blocks preventing proper message merging.
96- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
97- Fixed an issue where a single file write/edit error would abort all other parallel file write/edit operations. Independent file mutations now complete even when a sibling fails.
98- Added `last_assistant_message` field to Stop and SubagentStop hook inputs, providing the final assistant response text so hooks can access it without parsing transcript files.
99- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
100- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
101- Fixed an issue where bash commands with backslash-newline continuation lines (e.g., long commands split across multiple lines with `\`) would produce spurious empty arguments, potentially breaking command execution.
102- Fixed built-in slash commands (`/help`, `/model`, `/compact`, etc.) being hidden from the autocomplete dropdown when many user skills are installed (anthropics/claude-code#22020)
103- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
104- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
105- Fixed crash when a skill's `name` or `description` in SKILL.md frontmatter is a bare number (e.g., `name: 3000`) — the value is now properly coerced to a string (anthropics/claude-code#25837)
106- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
107- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
108- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
109- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
110- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
111- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
112- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
113- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
114- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
115- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
116- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
117- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
118- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
119- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
120- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
121- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
122- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
123- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
124- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
125- Fixed React crash (error #31) when a skill's `argument-hint` in SKILL.md frontmatter uses YAML sequence syntax (e.g., `[topic: foo | bar]`) — the value is now properly coerced to a string (anthropics/claude-code#25826)
126- Fixed crash when using `/fork` on sessions that used web search — null entries in search results from transcript deserialization are now handled gracefully (anthropics/claude-code#25811)
127- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
128- Fixed custom agents and skills not being discovered when running from a git worktree — project-level `.claude/agents/` and `.claude/skills/` from the main repository are now included (anthropics/claude-code#25816)
129- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
130- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
131- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
132- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
133- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
134- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
135- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
136- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
137- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
138- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
139- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
140- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
141- Fixed background tasks failing in git worktrees due to remote URL resolution reading from worktree-specific gitdir instead of the main repository config (anthropics/claude-code#26065)
142- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
143- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
144- Fixed Edit tool silently corrupting Unicode curly quotes (\u201c\u201d \u2018\u2019) by replacing them with straight quotes when making edits (anthropics/claude-code#26141)
145- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
146
147## 2.1.46
148
149- Fixed orphaned CC processes after terminal disconnect on macOS
150- Added support for using claude.ai MCP connectors in Claude Code
151
152## 2.1.45
153
154- Added support for Claude Sonnet 4.6
155- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
156- Added `spinnerTipsOverride` setting to customize spinner tips — configure `tips` with an array of custom tip strings, and optionally set `excludeDefault: true` to show only your custom tips instead of the built-in ones
157- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
158- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
159- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
160- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
161- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
162- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
163- Fixed excessive `.claude.json.backup` files accumulating on every startup
164- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
165- Improved startup performance by removing eager loading of session history for stats caching
166- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
167- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
168- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
169
170## 2.1.44
171
172- Fixed ENAMETOOLONG errors for deeply-nested directory paths
173- Fixed auth refresh errors
174
175## 2.1.43
176
177- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
178- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
179- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
180
181## 2.1.42
182
183- Improved startup performance by deferring Zod schema construction
184- Improved prompt cache hit rates by moving date out of system prompt
185- Added one-time Opus 4.6 effort callout for eligible users
186- Fixed /resume showing interrupt messages as session titles
187- Fixed image dimension limit errors to suggest /compact
188
189## 2.1.41
190
191- Added guard against launching Claude Code inside another Claude Code session
192- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
193- Fixed a crash when MCP tools return image content during streaming
194- Fixed /resume session previews showing raw XML tags instead of readable command names
195- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
196- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
197- Fixed hook blocking errors (exit code 2) not showing stderr to the user
198- Added `speed` attribute to OTel events and trace spans for fast mode visibility
199- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
200- Added Windows ARM64 (win32-arm64) native binary support
201- Improved `/rename` to auto-generate session name from conversation context when called without arguments
202- Improved narrow terminal layout for prompt footer
203- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
204- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
205- Fixed background task notifications not being delivered in streaming Agent SDK mode
206- Fixed cursor jumping to end on each keystroke in classifier rule input
207- Fixed markdown link display text being dropped for raw URL
208- Fixed auto-compact failure error notifications being shown to users
209- Fixed permission wait time being included in subagent elapsed time display
210- Fixed proactive ticks firing while in plan mode
211- Fixed clear stale permission rules when settings change on disk
212- Fixed hook blocking errors showing stderr content in UI
213
214## 2.1.39
215
216- Improved terminal rendering performance
217- Fixed fatal errors being swallowed instead of displayed
218- Fixed process hanging after session close
219- Fixed character loss at terminal screen boundary
220- Fixed blank lines in verbose transcript view
221
222## 2.1.38
223
224- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
225- Fixed Tab key queueing slash commands instead of autocompleting
226- Fixed bash permission matching for commands using environment variable wrappers
227- Fixed text between tool uses disappearing when not using streaming
228- Fixed duplicate sessions when resuming in VS Code extension
229- Improved heredoc delimiter parsing to prevent command smuggling
230- Blocked writes to `.claude/skills` directory in sandbox mode
231
232## 2.1.37
233
234- Fixed an issue where /fast was not immediately available after enabling /extra-usage
235
236## 2.1.36
237
238- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
239
240## 2.1.34
241
242- Fixed a crash when agent teams setting changed between renders
243- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
244
245## 2.1.33
246
247- Fixed agent teammate sessions in tmux to send and receive messages
248- Fixed warnings about agent teams not being available on your current plan
249- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
250- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
251- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
252- Added plugin name to skill descriptions and `/skills` menu for better discoverability
253- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
254- 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
255- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
256- 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
257- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
258- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
259- Errors from invalid managed settings are now surfaced
260- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
261- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
262- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
263
264## 2.1.32
265
266- Claude Opus 4.6 is now available!
267- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
268- Claude now automatically records and recalls memories as it works
269- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
270- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
271- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
272- Updated --resume to re-use --agent value specified in previous conversation by default.
273- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
274- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
275- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
276- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
277- VSCode: Added spinner when loading past conversations list
278
279## 2.1.31
280
281- Added session resume hint on exit, showing how to continue your conversation later
282- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
283- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
284- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
285- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
286- 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
287- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
288- 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
289- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
290- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
291- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
292
293## 2.1.30
294
295- 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.
296- 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`.
297- Added `/debug` for Claude to help troubleshoot the current session
298- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
299- Added token count, tool uses, and duration metrics to Task tool results
300- Added reduced motion mode to the config
301- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
302- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
303- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
304- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
305- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
306- Fixed permission dialogs stealing focus while actively typing
307- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
308- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
309- 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
310- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
311- Changed `/model` to execute immediately instead of being queued
312- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
313- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
314
315## 2.1.29
316
317- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
318
319## 2.1.27
320
321- Added tool call failures and denials to debug logs
322- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
323- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
324- Sessions are now automatically linked to PRs when created via `gh pr create`
325- Fixed /context command not displaying colored output
326- Fixed status bar duplicating background task indicator when PR status was shown
327- Windows: Fixed bash command execution failing for users with `.bashrc` files
328- Windows: Fixed console windows flashing when spawning child processes
329- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
330
331## 2.1.25
332
333- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
334
335## 2.1.23
336
337- Added customizable spinner verbs setting (`spinnerVerbs`)
338- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
339- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
340- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
341- Fixed pending async hooks not being cancelled when headless streaming sessions ended
342- Fixed tab completion not updating the input field when accepting a suggestion
343- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
344- Improved terminal rendering performance with optimized screen data layout
345- Changed Bash commands to show timeout duration alongside elapsed time
346- Changed merged pull requests to show a purple status indicator in the prompt footer
347- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
348
349## 2.1.22
350
351- Fixed structured outputs for non-interactive (-p) mode
352
353## 2.1.21
354
355- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
356- Fixed shell completion cache files being truncated on exit
357- Fixed API errors when resuming sessions that were interrupted during tool execution
358- Fixed auto-compact triggering too early on models with large output token limits
359- Fixed task IDs potentially being reused after deletion
360- Fixed file search not working in VS Code extension on Windows
361- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
362- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
363- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
364- [VSCode] Fixed message action buttons having incorrect background colors
365
366## 2.1.20
367
368- Added arrow key history navigation in vim normal mode when cursor cannot move further
369- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
370- 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
371- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
372- Added ability to delete tasks via the `TaskUpdate` tool
373- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
374- Fixed agents sometimes ignoring user messages sent while actively working on a task
375- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
376- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
377- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
378- Fixed draft prompt being lost when pressing UP arrow to navigate command history
379- Fixed ghost text flickering when typing slash commands mid-input
380- Fixed marketplace source removal not properly deleting settings
381- Fixed duplicate output in some commands like `/context`
382- Fixed task list sometimes showing outside the main conversation view
383- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
384- Fixed crashes when cancelling tool use
385- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
386- Improved thinking status text with a subtle shimmer animation
387- Improved task list to dynamically adjust visible items based on terminal height
388- Improved fork conversation hint to show how to resume the original session
389- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
390- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
391- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
392- Changed the `/copy` command to be available to all users
393- Changed background agents to prompt for tool permissions before launching
394- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
395- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
396
397## 2.1.19
398
399- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
400- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
401- Fixed crashes on processors without AVX instruction support
402- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
403- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
404- Fixed resuming sessions by custom title not working when run from a different directory
405- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
406- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
407- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
408- Fixed file write preview omitting empty lines
409- Changed skills without additional permissions or hooks to be allowed without requiring approval
410- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
411- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
412- [VSCode] Enabled session forking and rewind functionality for all users
413
414## 2.1.18
415
416- 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
417
418## 2.1.17
419
420- Fixed crashes on processors without AVX instruction support
421
422## 2.1.16
423
424- Added new task management system, including new capabilities like dependency tracking
425- [VSCode] Added native plugin management support
426- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
427- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
428- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
429- Fixed session titles on the resume screen not respecting the user's language setting
430- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
431
432## 2.1.15
433
434- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
435- Improved UI rendering performance with React Compiler
436- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
437- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
438
439## 2.1.14
440
441- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
442- Added search to installed plugins list - type to filter by name or description
443- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
444- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
445- Fixed memory issues that could cause crashes when running parallel subagents
446- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
447- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
448- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
449- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
450- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
451- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
452- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
453- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
454- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
455- Improved backspace to delete pasted text as a single token instead of one character at a time
456- [VSCode] Added `/usage` command to display current plan usage
457
458## 2.1.12
459
460- Fixed message rendering bug
461
462## 2.1.11
463
464- Fixed excessive MCP connection requests for HTTP/SSE transports
465
466## 2.1.10
467
468- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
469- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
470- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
471- Improved startup to capture keystrokes typed before the REPL is fully ready
472- Improved file suggestions to show as removable attachments instead of inserting text when accepted
473- [VSCode] Added install count display to plugin listings
474- [VSCode] Added trust warning when installing plugins
475
476## 2.1.9
477
478- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
479- Added `plansDirectory` setting to customize where plan files are stored
480- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
481- Added session URL attribution to commits and PRs created from web sessions
482- Added support for `PreToolUse` hooks to return `additionalContext` to the model
483- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
484- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
485- Fixed MCP server reconnection hanging when cached connection promise never resolves
486- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
487
488## 2.1.7
489
490- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
491- Added ability to provide feedback when accepting permission prompts
492- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
493- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
494- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
495- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
496- 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)
497- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
498- Fixed terminal title animation jitter by using fixed-width braille characters
499- Fixed plugins with git submodules not being fully initialized when installed
500- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
501- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
502- 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.
503- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
504- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
505
506## 2.1.6
507
508- Added search functionality to `/config` command for quickly filtering settings
509- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
510- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
511- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
512- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
513- Added an error display when the editor fails during Ctrl+G
514- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
515- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
516- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
517- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
518- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
519- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
520- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
521- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
522- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
523- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
524- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
525- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
526- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
527- Improved @ autocomplete with icons for different suggestion types and single-line formatting
528- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
529- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
530- Changed terminal title to "Claude Code" on startup for better window identification
531- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
532- [VSCode] Fixed usage indicator not updating after manual compact
533
534## 2.1.5
535
536- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
537
538## 2.1.4
539
540- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
541- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
542
543## 2.1.3
544
545- Merged slash commands and skills, simplifying the mental model with no change in behavior
546- Added release channel (`stable` or `latest`) toggle to `/config`
547- 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
548- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
549- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
550- Fixed mismatch between background task count in status bar and items shown in tasks dialog
551- Fixed sub-agents using the wrong model during conversation compaction
552- Fixed web search in sub-agents using incorrect model
553- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
554- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
555- Improved slash command suggestion readability by truncating long descriptions to 2 lines
556- Changed tool hook execution timeout from 60 seconds to 10 minutes
557- [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)
558
559## 2.1.2
560
561- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
562- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
563- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
564- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
565- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
566- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
567- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
568- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
569- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
570- Fixed updates incorrectly claiming another installation is in progress
571- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
572- Fixed remote session URL and teleport being broken when using `/tasks` command
573- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
574- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
575- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
576- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
577- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
578- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
579- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
580- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
581- [SDK] Changed minimum zod peer dependency to ^4.0.0
582- [VSCode] Fixed usage display not updating after manual compact
583
584## 2.1.0
585
586- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
587- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
588- Added support for `agent` field in skills to specify agent type for execution
589- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
590- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
591- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
592- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
593- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
594- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
595- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
596- Fixed Esc key with queued prompts to only move them to input without canceling the running task
597- Reduced permission prompts for complex bash commands
598- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
599- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
600- Fixed plugin path resolution for file-based marketplace sources
601- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
602- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
603- Fixed Claude in Chrome support for WSL environments
604- Fixed Windows native installer silently failing when executable creation fails
605- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
606- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
607- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
608- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
609- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
610- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
611- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
612- Added hooks support for skill and slash command frontmatter
613- 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
614- Added `/plan` command shortcut to enable plan mode directly from the prompt
615- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
616- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
617- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
618- Added support for `once: true` config for hooks
619- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
620- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
621- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
622- Added left/right arrow key navigation for cycling through tabs in dialogs
623- Added real-time thinking block display in Ctrl+O transcript mode
624- Added filepath to full output in background bash task details dialog
625- Added Skills as a separate category in the context visualization
626- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
627- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
628- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
629- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
630- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
631- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
632- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
633- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
634- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
635- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
636- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
637- Fixed commands with `$()` command substitution failing with parse errors
638- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
639- 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)
640- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
641- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
642- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
643- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
644- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
645- Fixed subagents sometimes not inheriting the parent's model by default
646- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
647- Fixed duplicate Bash commands appearing in permission request option labels
648- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
649- Fixed background task completion notifications to appear proactively with bullet point
650- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
651- Fixed cursor disappearing after dismissing permission dialogs
652- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
653- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
654- Fixed images being silently dropped when queueing messages while backgrounding a task
655- Fixed large pasted images failing with "Image was too large" error
656- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
657- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
658- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
659- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
660- Fixed spinner token counter to properly accumulate tokens from subagents during execution
661- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
662- Fixed race condition where LSP tool could return "no server available" during startup
663- Fixed feedback submission hanging indefinitely when network requests timeout
664- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
665- Fixed hook success message showing trailing colon when hook has no output
666- Multiple optimizations to improve startup performance
667- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
668- Improved performance when reading Jupyter notebooks with many cells
669- Improved reliability for piped input like `cat refactor.md | claude`
670- Improved reliability for AskQuestion tool
671- Improved sed in-place edit commands to render as file edits with diff preview
672- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
673- Improved compaction reliability
674- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
675- Improved skills to show progress while executing, displaying tool uses as they happen
676- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
677- Improved skill suggestions to prioritize recently and frequently used skills
678- Improved spinner feedback when waiting for the first response token
679- Improved token count display in spinner to include tokens from background agents
680- Improved incremental output for async agents to give the main thread more control and visibility
681- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
682- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
683- Improved macOS screenshot paste reliability with TIFF format support
684- Improved `/stats` output
685- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
686- Changed "Interrupted" message color from red to grey for a less alarming appearance
687- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
688- Removed underline styling from image reference links
689- [SDK] Changed minimum zod peer dependency to ^4.0.0
690- [VSCode] Added currently selected model name to the context menu
691- [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")
692- [VSCode] Fixed paragraph breaks not rendering in markdown content
693- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
694- [Windows] Fixed issue with improper rendering
695
696## 2.0.76
697
698- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
699
700## 2.0.75
701
702- Minor bugfixes
703
704## 2.0.74
705
706- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
707- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
708- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
709- Added syntax highlighting info to theme picker
710- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
711- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
712- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
713- Fixed a potential crash when syntax highlighting isn't initialized correctly
714- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
715- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
716- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
717- [Windows] Fixed issue with improper rendering
718- [VSCode] Added gift tag pictogram for year-end promotion message
719
720## 2.0.73
721
722- Added clickable `[Image #N]` links that open attached images in the default viewer
723- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
724- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
725- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
726- Fixed slow input history cycling and race condition that could overwrite text after message submission
727- Improved `/theme` command to open theme picker directly
728- Improved theme picker UI
729- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
730- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
731
732## 2.0.72
733
734- 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
735- Reduced terminal flickering
736- Added scannable QR code to mobile app tip for quick app downloads
737- Added loading indicator when resuming conversations for better feedback
738- Fixed `/context` command not respecting custom system prompts in non-interactive mode
739- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
740- Improved @ mention file suggestion speed (~3× faster in git repositories)
741- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
742- Improved settings validation errors to be more prominent
743- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
744
745## 2.0.71
746
747- Added /config toggle to enable/disable prompt suggestions
748- Added `/settings` as an alias for the `/config` command
749- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
750- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
751- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
752- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
753- New syntax highlighting engine for native build
754
755## 2.0.70
756
757- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
758- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
759- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
760- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
761- Fixed input being cleared when processing queued commands while the user was typing
762- Fixed prompt suggestions replacing typed input when pressing Tab
763- Fixed diff view not updating when terminal is resized
764- Improved memory usage by 3x for large conversations
765- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
766- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
767- Fix thinking mode toggle in /config not persisting correctly
768- Improve UI for file creation permission dialog
769
770## 2.0.69
771
772- Minor bugfixes
773
774## 2.0.68
775
776- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
777- Fixed a bug where disallowed MCP tools were visible to the model
778- Fixed an issue where steering messages could be lost while a subagent is working
779- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
780- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
781- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
782
783## 2.0.67
784
785- Thinking mode is now enabled by default for Opus 4.5
786- Thinking mode configuration has moved to /config
787- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
788- Show reason why autoupdater is disabled in `/doctor`
789- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
790- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
791- Fixed scroll position resetting after deleting a permission rule in `/permissions`
792- 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
793- Fixed `claude install --force` not bypassing stale lock files
794- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
795- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
796
797## 2.0.65
798
799- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
800- Added context window information to status line input
801- Added `fileSuggestion` setting for custom `@` file search commands
802- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
803- Fixed prompt not being saved to history when aborting a query with Escape
804- Fixed Read tool image handling to identify format from bytes instead of file extension
805
806## 2.0.64
807
808- Made auto-compacting instant
809- Agents and bash commands can run asynchronously and send messages to wake up the main agent
810- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
811- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
812- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
813- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
814- Fixed auto-loading .env when using native installer
815- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
816- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
817- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
818- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
819- Bedrock: Improve efficiency of token counting
820- Bedrock: Add support for `aws login` AWS Management Console credentials
821- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
822
823## 2.0.62
824
825- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
826- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
827- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
828- Fixed slash command selection not working when multiple commands share the same name
829- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
830- Fixed running versions getting removed because lock file incorrectly going stale
831- Fixed IDE diff tab not closing when rejecting file changes
832
833## 2.0.61
834
835- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
836
837## 2.0.60
838
839- Added background agent support. Agents run in the background while you work
840- Added --disable-slash-commands CLI flag to disable all slash commands
841- Added model name to "Co-Authored-By" commit messages
842- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
843- Updated Fetch to skip summarization for pre-approved websites
844- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
845
846## 2.0.59
847
848- Added --agent CLI flag to override the agent setting for the current session
849- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
850- VS Code: Fixed .claude.json config file being read from incorrect location
851
852## 2.0.58
853
854- Pro users now have access to Opus 4.5 as part of their subscription!
855- Fixed timer duration showing "11m 60s" instead of "12m 0s"
856- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
857
858## 2.0.57
859
860- Added feedback input when rejecting plans, allowing users to tell Claude what to change
861- VSCode: Added streaming message support for real-time response display
862
863## 2.0.56
864
865- Added setting to enable/disable terminal progress bar (OSC 9;4)
866- 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.
867
868## 2.0.55
869
870- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
871- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
872- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
873- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
874
875## 2.0.54
876
877- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
878- Fix issue with excessive iTerm notifications
879
880## 2.0.52
881
882- Fixed duplicate message display when starting Claude with a command line argument
883- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
884- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
885- Permit some uses of `$!` in bash commands
886
887## 2.0.51
888
889- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
890- Introducing Claude Code for Desktop: https://claude.com/download
891- 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
892- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
893- Plan Mode now builds more precise plans and executes more thoroughly
894- Usage limit notifications now easier to understand
895- Switched `/usage` back to "% used"
896- Fixed handling of thinking errors
897- Fixed performance regression
898
899## 2.0.50
900
901- Fixed bug preventing calling MCP tools that have nested references in their input schemas
902- Silenced a noisy but harmless error during upgrades
903- Improved ultrathink text display
904- Improved clarity of 5-hour session limit warning message
905
906## 2.0.49
907
908- Added readline-style ctrl-y for pasting deleted text
909- Improved clarity of usage limit warning message
910- Fixed handling of subagent permissions
911
912## 2.0.47
913
914- Improved error messages and validation for `claude --teleport`
915- Improved error handling in `/usage`
916- Fixed race condition with history entry not getting logged at exit
917- Fixed Vertex AI configuration not being applied from `settings.json`
918
919## 2.0.46
920
921- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
922
923## 2.0.45
924
925- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
926- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
927- Send background tasks to Claude Code on the web by starting a message with `&`
928
929## 2.0.43
930
931- Added `permissionMode` field for custom agents
932- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
933- Added skills frontmatter field to declare skills to auto-load for subagents
934- Added the `SubagentStart` hook event
935- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
936- Fixed duplicate rendering of some messages in the UI
937- Fixed some visual flickers
938- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
939
940## 2.0.42
941
942- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
943
944## 2.0.41
945
946- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
947- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
948- Fixed incorrect labeling of user settings vs project settings in command descriptions
949- Fixed crash when plugin command hooks timeout during execution
950- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
951- Fixed broken security documentation links in trust dialogs and onboarding
952- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
953- ctrl-r history search landing on a slash command no longer cancels the search
954- SDK: Support custom timeouts for hooks
955- Allow more safe git commands to run without approval
956- Plugins: Added support for sharing and installing output styles
957- Teleporting a session from web will automatically set the upstream branch
958
959## 2.0.37
960
961- Fixed how idleness is computed for notifications
962- Hooks: Added matcher values for Notification hook events
963- Output Styles: Added `keep-coding-instructions` option to frontmatter
964
965## 2.0.36
966
967- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
968- Fixed queued messages being incorrectly executed as bash commands
969- Fixed input being lost when typing while a queued message is processed
970
971## 2.0.35
972
973- Improve fuzzy search results when searching commands
974- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
975- 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
976- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
977- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
978
979## 2.0.34
980
981- VSCode Extension: Added setting to configure the initial permission mode for new conversations
982- Improved file path suggestion performance with native Rust-based fuzzy finder
983- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
984- Fixed memory crash when reading or writing large files (especially base64-encoded images)
985
986## 2.0.33
987
988- Native binary installs now launch quicker.
989- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
990- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
991
992## 2.0.32
993
994- Un-deprecate output styles based on community feedback
995- Added `companyAnnouncements` setting for displaying announcements on startup
996- Fixed hook progress messages not updating correctly during PostToolUse hook execution
997
998## 2.0.31
999
1000- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1001- Vertex: add support for Web Search on supported models
1002- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1003- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1004- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1005- Fixed plugin uninstall not removing plugins
1006
1007## 2.0.30
1008
1009- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1010- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1011- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1012- Added prompt-based stop hooks
1013- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1014- Enabled SSE MCP servers on native build
1015- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1016- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1017- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1018- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1019- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1020- Fixed bug that saved session permissions to local settings
1021- Fixed MCP tools not being available to sub-agents
1022- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1023- Fixed delay when navigating through typeahead suggestions with arrow keys
1024- VSCode: Restored selection indicator in input footer showing current file or code selection status
1025
1026## 2.0.28
1027
1028- Plan mode: introduced new Plan subagent
1029- Subagents: claude can now choose to resume subagents
1030- Subagents: claude can dynamically choose the model used by its subagents
1031- SDK: added --max-budget-usd flag
1032- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1033- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1034- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1035- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1036- Various other bug fixes
1037
1038## 2.0.27
1039
1040- New UI for permission prompts
1041- Added current branch filtering and search to session resume screen for easier navigation
1042- Fixed directory @-mention causing "No assistant message found" error
1043- VSCode Extension: Add config setting to include .gitignored files in file searches
1044- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1045
1046## 2.0.25
1047
1048- 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
1049
1050## 2.0.24
1051
1052- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1053- Claude Code Web: Support for Web -> CLI teleport
1054- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1055- Bedrock: Display awsAuthRefresh output when auth is required
1056
1057## 2.0.22
1058
1059- Fixed content layout shift when scrolling through slash commands
1060- IDE: Add toggle to enable/disable thinking.
1061- Fix bug causing duplicate permission prompts with parallel tool calls
1062- Add support for enterprise managed MCP allowlist and denylist
1063
1064## 2.0.21
1065
1066- Support MCP `structuredContent` field in tool responses
1067- Added an interactive question tool
1068- Claude will now ask you questions more often in plan mode
1069- Added Haiku 4.5 as a model option for Pro users
1070- Fixed an issue where queued commands don't have access to previous messages' output
1071
1072## 2.0.20
1073
1074- Added support for Claude Skills
1075
1076## 2.0.19
1077
1078- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1079- Fixed a bug where Haiku was unnecessarily called in print mode
1080
1081## 2.0.17
1082
1083- Added Haiku 4.5 to model selector!
1084- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1085- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1086- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1087- OTEL: support HTTP_PROXY and HTTPS_PROXY
1088- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1089
1090## 2.0.15
1091
1092- Fixed bug with resuming where previously created files needed to be read again before writing
1093- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1094
1095## 2.0.14
1096
1097- Fix @-mentioning MCP servers to toggle them on/off
1098- Improve permission checks for bash with inline env vars
1099- Fix ultrathink + thinking toggle
1100- Reduce unnecessary logins
1101- Document --system-prompt
1102- Several improvements to rendering
1103- Plugins UI polish
1104
1105## 2.0.13
1106
1107- Fixed `/plugin` not working on native build
1108
1109## 2.0.12
1110
1111- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1112- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1113- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1114- `/plugin validate` command for validating plugin structure and configuration
1115- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1116- Plugin documentation available at https://code.claude.com/docs/en/plugins
1117- Comprehensive error messages and diagnostics via `/doctor` command
1118- Avoid flickering in `/model` selector
1119- Improvements to `/help`
1120- Avoid mentioning hooks in `/resume` summaries
1121- Changes to the "verbose" setting in `/config` now persist across sessions
1122
1123## 2.0.11
1124
1125- Reduced system prompt size by 1.4k tokens
1126- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1127- Fixed Opus fallback rate limit errors appearing incorrectly
1128- Fixed /add-dir command selecting wrong default tab
1129
1130## 2.0.10
1131
1132- Rewrote terminal renderer for buttery smooth UI
1133- Enable/disable MCP servers by @mentioning, or in /mcp
1134- Added tab completion for shell commands in bash mode
1135- PreToolUse hooks can now modify tool inputs
1136- Press Ctrl-G to edit your prompt in your system's configured text editor
1137- Fixes for bash permission checks with environment variables in the command
1138
1139## 2.0.9
1140
1141- Fix regression where bash backgrounding stopped working
1142
1143## 2.0.8
1144
1145- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1146- IDE: Add drag-and-drop support for files and folders in chat
1147- /context: Fix counting for thinking blocks
1148- Improve message rendering for users with light themes on dark terminals
1149- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1150
1151## 2.0.5
1152
1153- IDE: Fix IME unintended message submission with Enter and Tab
1154- IDE: Add "Open in Terminal" link in login screen
1155- Fix unhandled OAuth expiration 401 API errors
1156- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1157
1158## 2.0.1
1159
1160- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1161- Various bug fixes and presentation improvements
1162
1163## 2.0.0
1164
1165- New native VS Code extension
1166- Fresh coat of paint throughout the whole app
1167- /rewind a conversation to undo code changes
1168- /usage command to see plan limits
1169- Tab to toggle thinking (sticky across sessions)
1170- Ctrl-R to search history
1171- Unshipped claude config command
1172- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1173- SDK: The Claude Code SDK is now the Claude Agent SDK
1174- Add subagents dynamically with `--agents` flag
1175
1176## 1.0.126
1177
1178- Enable /context command for Bedrock and Vertex
1179- Add mTLS support for HTTP-based OpenTelemetry exporters
1180
1181## 1.0.124
1182
1183- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1184- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1185- No longer inform Claude of the list of allowed tools when permission is denied
1186- Fixed security vulnerability in Bash tool permission checks
1187- Improved VSCode extension performance for large files
1188
1189## 1.0.123
1190
1191- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1192- Fixed thinking mode triggering on negation phrases like "don't think"
1193- Fixed rendering performance degradation during token streaming
1194- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1195- Enhanced BashTool environment snapshot logging
1196- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1197- Migrated --debug logging to a file, to enable easy tailing & filtering
1198
1199## 1.0.120
1200
1201- Fix input lag during typing, especially noticeable with large prompts
1202- Improved VSCode extension command registry and sessions dialog user experience
1203- Enhanced sessions dialog responsiveness and visual feedback
1204- Fixed IDE compatibility issue by removing worktree support check
1205- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1206
1207## 1.0.119
1208
1209- Fix Windows issue where process visually freezes on entering interactive mode
1210- Support dynamic headers for MCP servers via headersHelper configuration
1211- Fix thinking mode not working in headless sessions
1212- Fix slash commands now properly update allowed tools instead of replacing them
1213
1214## 1.0.117
1215
1216- Add Ctrl-R history search to recall previous commands like bash/zsh
1217- Fix input lag while typing, especially on Windows
1218- Add sed command to auto-allowed commands in acceptEdits mode
1219- Fix Windows PATH comparison to be case-insensitive for drive letters
1220- Add permissions management hint to /add-dir output
1221
1222## 1.0.115
1223
1224- Improve thinking mode display with enhanced visual effects
1225- Type /t to temporarily disable thinking mode in your prompt
1226- Improve path validation for glob and grep tools
1227- Show condensed output for post-tool hooks to reduce visual clutter
1228- Fix visual feedback when loading state completes
1229- Improve UI consistency for permission request dialogs
1230
1231## 1.0.113
1232
1233- Deprecated piped input in interactive mode
1234- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1235
1236## 1.0.112
1237
1238- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1239- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1240- Hooks: Added systemMessage support for SessionEnd hooks
1241- Added `spinnerTipsEnabled` setting to disable spinner tips
1242- IDE: Various improvements and bug fixes
1243
1244## 1.0.111
1245
1246- /model now validates provided model names
1247- Fixed Bash tool crashes caused by malformed shell syntax parsing
1248
1249## 1.0.110
1250
1251- /terminal-setup command now supports WezTerm
1252- MCP: OAuth tokens now proactively refresh before expiration
1253- Fixed reliability issues with background Bash processes
1254
1255## 1.0.109
1256
1257- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1258
1259## 1.0.106
1260
1261- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1262
1263## 1.0.97
1264
1265- Settings: /doctor now validates permission rule syntax and suggests corrections
1266
1267## 1.0.94
1268
1269- Vertex: add support for global endpoints for supported models
1270- /memory command now allows direct editing of all imported memory files
1271- SDK: Add custom tools as callbacks
1272- Added /todos command to list current todo items
1273
1274## 1.0.93
1275
1276- Windows: Add alt + v shortcut for pasting images from clipboard
1277- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1278
1279## 1.0.90
1280
1281- Settings file changes take effect immediately - no restart required
1282
1283## 1.0.88
1284
1285- Fixed issue causing "OAuth authentication is currently not supported"
1286- Status line input now includes `exceeds_200k_tokens`
1287- Fixed incorrect usage tracking in /cost.
1288- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1289- Bedrock: Updated default Sonnet model to Sonnet 4
1290
1291## 1.0.86
1292
1293- Added /context to help users self-serve debug context issues
1294- SDK: Added UUID support for all SDK messages
1295- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1296
1297## 1.0.85
1298
1299- Status line input now includes session cost info
1300- Hooks: Introduced SessionEnd hook
1301
1302## 1.0.84
1303
1304- Fix tool_use/tool_result id mismatch error when network is unstable
1305- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1306- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1307- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1308
1309## 1.0.83
1310
1311- @-mention: Support files with spaces in path
1312- New shimmering spinner
1313
1314## 1.0.82
1315
1316- SDK: Add request cancellation support
1317- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1318- Settings: Validation prevents invalid fields in .claude/settings.json files
1319- MCP: Improve tool name consistency
1320- Bash: Fix crash when Claude tries to automatically read large files
1321
1322## 1.0.81
1323
1324- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1325- Agents: Fix custom agent loading when agent files are unparsable
1326
1327## 1.0.80
1328
1329- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1330
1331## 1.0.77
1332
1333- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1334- SDK: Add session support and permission denial tracking
1335- Fix token limit errors in conversation summarization
1336- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1337
1338## 1.0.73
1339
1340- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1341- MCP: Press Esc to cancel OAuth authentication flows
1342- Bash: Improved command validation and reduced false security warnings
1343- UI: Enhanced spinner animations and status line visual hierarchy
1344- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1345
1346## 1.0.72
1347
1348- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1349
1350## 1.0.71
1351
1352- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1353- Customizable status line: add your terminal prompt to Claude Code with /statusline
1354
1355## 1.0.70
1356
1357- Performance: Optimized message rendering for better performance with large contexts
1358- Windows: Fixed native file search, ripgrep, and subagent functionality
1359- Added support for @-mentions in slash command arguments
1360
1361## 1.0.69
1362
1363- Upgraded Opus to version 4.1
1364
1365## 1.0.68
1366
1367- Fix incorrect model names being used for certain commands like `/pr-comments`
1368- 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.
1369- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1370- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1371- SDK: Added canUseTool callback support for tool confirmation
1372- Added `disableAllHooks` setting
1373- Improved file suggestions performance in large repos
1374
1375## 1.0.65
1376
1377- IDE: Fixed connection stability issues and error handling for diagnostics
1378- Windows: Fixed shell environment setup for users without .bashrc files
1379
1380## 1.0.64
1381
1382- Agents: Added model customization support - you can now specify which model an agent should use
1383- Agents: Fixed unintended access to the recursive agent tool
1384- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1385- SDK: Fixed user input tracking across multi-turn conversations
1386- Added hidden files to file search and @-mention suggestions
1387
1388## 1.0.63
1389
1390- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1391
1392## 1.0.62
1393
1394- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1395- Hooks: Added SessionStart hook for new session initialization
1396- /add-dir command now supports typeahead for directory paths
1397- Improved network connectivity check reliability
1398
1399## 1.0.61
1400
1401- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1402- Settings: Added `--settings` flag to load settings from a JSON file
1403- Settings: Fixed resolution of settings files paths that are symlinks
1404- OTEL: Fixed reporting of wrong organization after authentication changes
1405- Slash commands: Fixed permissions checking for allowed-tools with Bash
1406- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1407- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1408- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1409
1410## 1.0.60
1411
1412- You can now create custom subagents for specialized tasks! Run /agents to get started
1413
1414## 1.0.59
1415
1416- SDK: Added tool confirmation support with canUseTool callback
1417- SDK: Allow specifying env for spawned process
1418- Hooks: Exposed PermissionDecision to hooks (including "ask")
1419- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1420- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1421
1422## 1.0.58
1423
1424- Added support for reading PDFs
1425- MCP: Improved server health status display in 'claude mcp list'
1426- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1427
1428## 1.0.57
1429
1430- Added support for specifying a model in slash commands
1431- Improved permission messages to help Claude understand allowed tools
1432- Fix: Remove trailing newlines from bash output in terminal wrapping
1433
1434## 1.0.56
1435
1436- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1437- Fixes for WSL IDE detection
1438- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1439
1440## 1.0.55
1441
1442- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1443- Windows: fixed Ctrl+Z crash
1444- SDK: Added ability to capture error logging
1445- Add --system-prompt-file option to override system prompt in print mode
1446
1447## 1.0.54
1448
1449- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1450- Custom slash commands: Added argument-hint to frontmatter
1451- Windows: OAuth uses port 45454 and properly constructs browser URL
1452- Windows: mode switching now uses alt + m, and plan mode renders properly
1453- Shell: Switch to in-memory shell snapshot to fix file-related errors
1454
1455## 1.0.53
1456
1457- Updated @-mention file truncation from 100 lines to 2000 lines
1458- 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).
1459
1460## 1.0.52
1461
1462- Added support for MCP server instructions
1463
1464## 1.0.51
1465
1466- Added support for native Windows (requires Git for Windows)
1467- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1468- Settings: /doctor can now help you identify and fix invalid setting files
1469- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1470- Increased auto-compact warning threshold from 60% to 80%
1471- Fixed an issue with handling user directories with spaces for shell snapshots
1472- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1473- Custom slash commands: Fixed user-level commands in subdirectories
1474- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1475
1476## 1.0.48
1477
1478- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1479- Added progress messages to Bash tool based on the last 5 lines of command output
1480- Added expanding variables support for MCP server configuration
1481- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1482- Improved IDE extension path handling when Claude Code runs in WSL
1483- Hooks: Added a PreCompact hook
1484- Vim mode: Added c, f/F, t/T
1485
1486## 1.0.45
1487
1488- Redesigned Search (Grep) tool with new tool input parameters and features
1489- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1490- Fixed config file corruption issue by enforcing atomic writes
1491- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1492- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1493- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1494
1495## 1.0.44
1496
1497- New /export command lets you quickly export a conversation for sharing
1498- MCP: resource_link tool results are now supported
1499- MCP: tool annotations and tool titles now display in /mcp view
1500- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1501
1502## 1.0.43
1503
1504- Fixed a bug where the theme selector was saving excessively
1505- Hooks: Added EPIPE system error handling
1506
1507## 1.0.42
1508
1509- Added tilde (`~`) expansion support to `/add-dir` command
1510
1511## 1.0.41
1512
1513- Hooks: Split Stop hook triggering into Stop and SubagentStop
1514- Hooks: Enabled optional timeout configuration for each command
1515- Hooks: Added "hook_event_name" to hook input
1516- Fixed a bug where MCP tools would display twice in tool list
1517- New tool parameters JSON for Bash tool in `tool_decision` event
1518
1519## 1.0.40
1520
1521- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1522
1523## 1.0.39
1524
1525- New Active Time metric in OpenTelemetry logging
1526
1527## 1.0.38
1528
1529- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1530
1531## 1.0.37
1532
1533- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1534
1535## 1.0.36
1536
1537- Web search now takes today's date into context
1538- Fixed a bug where stdio MCP servers were not terminating properly on exit
1539
1540## 1.0.35
1541
1542- Added support for MCP OAuth Authorization Server discovery
1543
1544## 1.0.34
1545
1546- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1547
1548## 1.0.33
1549
1550- Improved logging functionality with session ID support
1551- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1552- Improvements to plan mode
1553
1554## 1.0.32
1555
1556- Updated loopback config for litellm
1557- Added forceLoginMethod setting to bypass login selection screen
1558
1559## 1.0.31
1560
1561- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1562
1563## 1.0.30
1564
1565- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1566- Improved file path autocomplete with filename matching
1567- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1568- Enhanced jq regex support for complex filters with pipes and select
1569
1570## 1.0.29
1571
1572- Improved CJK character support in cursor navigation and rendering
1573
1574## 1.0.28
1575
1576- Slash commands: Fix selector display during history navigation
1577- Resizes images before upload to prevent API size limit errors
1578- Added XDG_CONFIG_HOME support to configuration directory
1579- Performance optimizations for memory usage
1580- New attributes (terminal.type, language) in OpenTelemetry logging
1581
1582## 1.0.27
1583
1584- Streamable HTTP MCP servers are now supported
1585- Remote MCP servers (SSE and HTTP) now support OAuth
1586- MCP resources can now be @-mentioned
1587- /resume slash command to switch conversations within Claude Code
1588
1589## 1.0.25
1590
1591- Slash commands: moved "project" and "user" prefixes to descriptions
1592- Slash commands: improved reliability for command discovery
1593- Improved support for Ghostty
1594- Improved web search reliability
1595
1596## 1.0.24
1597
1598- Improved /mcp output
1599- Fixed a bug where settings arrays got overwritten instead of merged
1600
1601## 1.0.23
1602
1603- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1604- Released Python SDK: pip install claude-code-sdk to get started
1605
1606## 1.0.22
1607
1608- SDK: Renamed `total_cost` to `total_cost_usd`
1609
1610## 1.0.21
1611
1612- Improved editing of files with tab-based indentation
1613- Fix for tool_use without matching tool_result errors
1614- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1615
1616## 1.0.18
1617
1618- Added --add-dir CLI argument for specifying additional working directories
1619- Added streaming input support without require -p flag
1620- Improved startup performance and session storage performance
1621- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1622- Added detailed MCP server tools display (/mcp)
1623- MCP authentication and permission improvements
1624- Added auto-reconnection for MCP SSE connections on disconnect
1625- Fixed issue where pasted content was lost when dialogs appeared
1626
1627## 1.0.17
1628
1629- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1630- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1631- MCP server list UI improvements
1632- Update Claude Code process title to display "claude" instead of "node"
1633
1634## 1.0.11
1635
1636- Claude Code can now also be used with a Claude Pro subscription
1637- Added /upgrade for smoother switching to Claude Max plans
1638- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1639- Improved shell configuration error handling
1640- Improved todo list handling during compaction
1641
1642## 1.0.10
1643
1644- Added markdown table support
1645- Improved streaming performance
1646
1647## 1.0.8
1648
1649- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1650- Increased default otel interval from 1s -> 5s
1651- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1652- Fixed a regression where search tools unnecessarily asked for permissions
1653- Added support for triggering thinking non-English languages
1654- Improved compacting UI
1655
1656## 1.0.7
1657
1658- Renamed /allowed-tools -> /permissions
1659- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1660- Deprecated claude config commands in favor of editing settings.json
1661- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1662- Improved error handling for /install-github-app
1663- Bugfixes, UI polish, and tool reliability improvements
1664
1665## 1.0.6
1666
1667- Improved edit reliability for tab-indented files
1668- Respect CLAUDE_CONFIG_DIR everywhere
1669- Reduced unnecessary tool permission prompts
1670- Added support for symlinks in @file typeahead
1671- Bugfixes, UI polish, and tool reliability improvements
1672
1673## 1.0.4
1674
1675- Fixed a bug where MCP tool errors weren't being parsed correctly
1676
1677## 1.0.1
1678
1679- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1680- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1681- Updated documentation links and OAuth process descriptions
1682
1683## 1.0.0
1684
1685- Claude Code is now generally available
1686- Introducing Sonnet 4 and Opus 4 models
1687
1688## 0.2.125
1689
1690- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1691- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1692
1693## 0.2.117
1694
1695- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1696- Introduced settings.cleanupPeriodDays
1697- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1698- Introduced --debug mode
1699
1700## 0.2.108
1701
1702- You can now send messages to Claude while it works to steer Claude in real-time
1703- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1704- Fixed a bug where thinking was not working in -p mode
1705- Fixed a regression in /cost reporting
1706- Deprecated MCP wizard interface in favor of other MCP commands
1707- Lots of other bugfixes and improvements
1708
1709## 0.2.107
1710
1711- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1712
1713## 0.2.106
1714
1715- MCP SSE server configs can now specify custom headers
1716- Fixed a bug where MCP permission prompt didn't always show correctly
1717
1718## 0.2.105
1719
1720- Claude can now search the web
1721- Moved system & account status to /status
1722- Added word movement keybindings for Vim
1723- Improved latency for startup, todo tool, and file edits
1724
1725## 0.2.102
1726
1727- Improved thinking triggering reliability
1728- Improved @mention reliability for images and folders
1729- You can now paste multiple large chunks into one prompt
1730
1731## 0.2.100
1732
1733- Fixed a crash caused by a stack overflow error
1734- Made db storage optional; missing db support disables --continue and --resume
1735
1736## 0.2.98
1737
1738- Fixed an issue where auto-compact was running twice
1739
1740## 0.2.96
1741
1742- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1743
1744## 0.2.93
1745
1746- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1747- Claude now has access to a Todo list that helps it stay on track and be more organized
1748
1749## 0.2.82
1750
1751- Added support for --disallowedTools
1752- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1753
1754## 0.2.75
1755
1756- Hit Enter to queue up additional messages while Claude is working
1757- Drag in or copy/paste image files directly into the prompt
1758- @-mention files to directly add them to context
1759- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1760- Improved performance for filename auto-complete
1761
1762## 0.2.74
1763
1764- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1765- Task tool can now perform writes and run bash commands
1766
1767## 0.2.72
1768
1769- Updated spinner to indicate tokens loaded and tool usage
1770
1771## 0.2.70
1772
1773- Network commands like curl are now available for Claude to use
1774- Claude can now run multiple web queries in parallel
1775- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1776
1777## 0.2.69
1778
1779- Fixed UI glitches with improved Select component behavior
1780- Enhanced terminal output display with better text truncation logic
1781
1782## 0.2.67
1783
1784- Shared project permission rules can be saved in .claude/settings.json
1785
1786## 0.2.66
1787
1788- Print mode (-p) now supports streaming output via --output-format=stream-json
1789- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1790
1791## 0.2.63
1792
1793- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1794
1795## 0.2.61
1796
1797- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1798- Enhanced image detection for more reliable clipboard paste functionality
1799- Fixed an issue where ESC key could crash the conversation history selector
1800
1801## 0.2.59
1802
1803- Copy+paste images directly into your prompt
1804- Improved progress indicators for bash and fetch tools
1805- Bugfixes for non-interactive mode (-p)
1806
1807## 0.2.54
1808
1809- Quickly add to Memory by starting your message with '#'
1810- Press ctrl+r to see full output for long tool results
1811- Added support for MCP SSE transport
1812
1813## 0.2.53
1814
1815- New web fetch tool lets Claude view URLs that you paste in
1816- Fixed a bug with JPEG detection
1817
1818## 0.2.50
1819
1820- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1821
1822## 0.2.49
1823
1824- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1825
1826## 0.2.47
1827
1828- Press Tab to auto-complete file and folder names
1829- Press Shift + Tab to toggle auto-accept for file edits
1830- Automatic conversation compaction for infinite conversation length (toggle with /config)
1831
1832## 0.2.44
1833
1834- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1835
1836## 0.2.41
1837
1838- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1839- MCP server startup no longer blocks the app from starting up
1840
1841## 0.2.37
1842
1843- New /release-notes command lets you view release notes at any time
1844- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1845
1846## 0.2.36
1847
1848- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1849- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1850
1851## 0.2.34
1852
1853- Vim bindings for text input - enable with /vim or /config
1854
1855## 0.2.32
1856
1857- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1858- Fix for some PersistentShell issues
1859
1860## 0.2.31
1861
1862- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1863- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1864
1865## 0.2.30
1866
1867- Added ANSI color theme for better terminal compatibility
1868- Fixed issue where slash command arguments weren't being sent properly
1869- (Mac-only) API keys are now stored in macOS Keychain
1870
1871## 0.2.26
1872
1873- New /approved-tools command for managing tool permissions
1874- Word-level diff display for improved code readability
1875- Fuzzy matching for slash commands
1876
1877## 0.2.21
1878
1879- Fuzzy matching for /commands