Skip to content
Developer Toolkit Developer Toolkit Developer Toolkit

Claude Code Version History

This comprehensive changelog tracks Claude Code’s rapid evolution since its preview launch in February 2025. Each version brings significant improvements based on user feedback and real-world usage patterns.

Version 2.1.x Series (Latest)

Section titled “Version 2.1.x Series (Latest)”

April 2026 Updates

Section titled “April 2026 Updates”

v2.1.119 - April 23, 2026

Persistent /config, Richer Telemetry & Security Polish

  • /config settings (theme, editor mode, verbose, etc.) now persist to ~/.claude/settings.json with proper project/local/policy override precedence
  • Added prUrlTemplate setting to point the footer PR badge at a custom code-review URL
  • Added CLAUDE_CODE_HIDE_CWD env var to hide the working directory in the startup logo
  • --from-pr now accepts GitLab merge-request, Bitbucket pull-request, and GitHub Enterprise PR URLs
  • --print mode honors agent tools:/disallowedTools: frontmatter; --agent <name> honors permissionMode
  • PowerShell tool commands can now be auto-approved in permission mode, matching Bash
  • Hooks: PostToolUse/PostToolUseFailure inputs now include duration_ms
  • OpenTelemetry: tool_result and tool_decision events include tool_use_id; tool_result also includes tool_input_size_bytes
  • Status line: stdin JSON now includes effort.level and thinking.enabled
  • Security: blockedMarketplaces now correctly enforces hostPattern and pathPattern entries
  • Subagent and SDK MCP server reconfiguration now connects in parallel instead of serially
  • Fixed pasting CRLF content (Windows clipboards, Xcode console) inserting an extra blank line
  • Fixed Glob/Grep tools disappearing on native builds when Bash is denied via permissions
  • Fixed scroll snap-back in fullscreen after every tool finish
  • Fixed MCP HTTP “Invalid OAuth error response” when servers return non-JSON OAuth discovery bodies
  • Fixed ${ENV_VAR} placeholders in MCP headers not substituted before requests

v2.1.118 - April 23, 2026

Vim Visual Mode, Named Themes & /usage

  • Added vim visual mode (v) and visual-line mode (V) with selection, operators, and visual feedback
  • Merged /cost and /stats into /usage — both remain as typing shortcuts
  • Custom themes: create and switch from /theme, or hand-edit ~/.claude/themes/; plugins can ship themes via a themes/ directory
  • Hooks can now invoke MCP tools directly via type: "mcp_tool"
  • Added DISABLE_UPDATES env var to block all update paths including manual claude update
  • WSL on Windows can inherit Windows-side managed settings via wslInheritsWindowsSettings
  • Auto mode: include "$defaults" in autoMode.allow/soft_deny/environment to add rules alongside built-ins
  • Added claude plugin tag to create release git tags for plugins with version validation
  • --continue/--resume now find sessions that added the current directory via /add-dir
  • Fixed /mcp menu hiding OAuth actions for servers with headersHelper
  • Fixed MCP servers whose OAuth response omits expires_in re-authenticating every hour
  • Fixed macOS keychain race overwriting freshly-refreshed OAuth tokens
  • Fixed credential save crash on Linux/Windows corrupting ~/.claude/.credentials.json
  • Fixed /fork writing full parent conversation per fork (now uses pointer + lazy hydrate)

v2.1.117 - April 22, 2026

Native bfs/ugrep on macOS/Linux, Parallel MCP Startup

  • Native macOS/Linux builds replace Glob/Grep tools with embedded bfs and ugrep available through the Bash tool — faster searches without the tool round-trip (Windows and npm builds unchanged)
  • Improved /model: selections persist across restarts even when the project pins a different model; startup header shows when the active model comes from a project/managed-settings pin
  • Faster startup when both local and claude.ai MCP servers are configured (concurrent connect default)
  • /resume on stale, large sessions now offers to summarize before re-reading
  • OpenTelemetry: user_prompt events include command_name/command_source; cost.usage, token.usage, api_request, api_error include effort when supported
  • Advisor Tool (experimental): dialog now carries an “experimental” label; no more stuck “result content could not be processed” errors
  • cleanupPeriodDays retention sweep now covers ~/.claude/tasks/, shell-snapshots/, and backups/
  • Default effort for Pro/Max on Opus 4.6 and Sonnet 4.6 is now high (was medium)
  • Fixed Opus 4.7 sessions showing inflated /context percentages — Claude Code was computing against 200K instead of native 1M
  • Fixed Bedrock application-inference-profile requests failing with 400 when backed by Opus 4.7 with thinking disabled
  • Fixed WebFetch hanging on very large HTML pages by truncating input before HTML→markdown conversion

v2.1.116 - April 20, 2026

Faster /resume, Thinking Spinner Progress

  • /resume on large sessions is significantly faster (up to 67% on 40MB+ sessions)
  • Faster MCP startup with multiple stdio servers; resources/templates/list is deferred to first @-mention
  • Smoother fullscreen scrolling in VS Code / Cursor / Windsurf terminals — /terminal-setup now configures editor scroll sensitivity
  • Thinking spinner shows progress inline (“still thinking”, “thinking more”, “almost done thinking”)
  • /config search now matches option values (e.g. “vim” finds Editor mode)
  • /doctor can now be opened while Claude is responding
  • Security: sandbox auto-allow no longer bypasses dangerous-path check for rm/rmdir on /, $HOME, or critical system directories
  • Claude Code now uses https://downloads.claude.ai/claude-code-releases
  • Fixed Devanagari and other Indic scripts rendering with broken column alignment
  • Fixed Ctrl+- not triggering undo in Kitty keyboard protocol terminals (iTerm2, Ghostty, WezTerm, Windows Terminal)
  • Fixed /branch rejecting conversations with transcripts larger than 50MB
  • Fixed intermittent API 400 “cache control TTL ordering” errors

v2.1.114 - April 18, 2026

Permission Dialog Crash Fix

  • Fixed a crash in the permission dialog when an agent teams teammate requested tool permission

v2.1.113 - April 17, 2026

Native CLI Binary, Sandbox Hardening

  • CLI now spawns a native Claude Code binary (via per-platform optional dependency) instead of bundled JavaScript
  • Added sandbox.network.deniedDomains to block specific domains even when a broader allowedDomains wildcard permits them
  • Fullscreen: Shift+↑/↓ scrolls the viewport when extending a selection past visible edge
  • Ctrl+A / Ctrl+E now move to start/end of current logical line in multiline input (readline behavior)
  • Windows: Ctrl+Backspace now deletes the previous word
  • Improved /loop: Esc cancels pending wakeups
  • Improved /ultrareview: faster launch, parallelized checks, diffstat in launch dialog
  • Subagents that stall mid-stream now fail with a clear error after 10 minutes instead of hanging
  • Security: macOS /private/{etc,var,tmp,home} paths treated as dangerous removal targets under Bash(rm:*)
  • Security: Bash deny rules now match commands wrapped in env/sudo/watch/ionice/setsid
  • Security: Bash(find:*) allow rules no longer auto-approve find -exec/-delete
  • Fixed thinking.type.enabled is not supported 400 error when using Opus 4.7 via Bedrock Application Inference Profile ARN
  • Fixed Bash dangerouslyDisableSandbox running commands outside the sandbox without a permission prompt

v2.1.112 - April 16, 2026

Opus 4.7 Availability Fix

  • Fixed “claude-opus-4-7 is temporarily unavailable” for auto mode

v2.1.111 - April 16, 2026

Claude Opus 4.7 + xhigh effort + /ultrareview

  • Claude Opus 4.7 xhigh effort level now available — tune speed vs. intelligence via /effort, --effort, or the model picker; other models fall back to high
  • Auto mode now available for Max subscribers when using Opus 4.7
  • /effort opens an interactive slider when called without arguments
  • Added “Auto (match terminal)” theme option that matches terminal dark/light mode
  • Added /less-permission-prompts skill — scans transcripts for read-only Bash/MCP calls and proposes a prioritized allowlist
  • Added /ultrareview for comprehensive multi-agent code review in the cloud — /ultrareview for current branch, /ultrareview <PR#> for a specific PR
  • Auto mode no longer requires --enable-auto-mode
  • Windows: PowerShell tool progressively rolling out; opt in/out with CLAUDE_CODE_USE_POWERSHELL_TOOL
  • Read-only bash with glob patterns (e.g. ls *.ts) and commands starting with cd <project-dir> && no longer prompt
  • Plan files now named after your prompt (e.g. fix-auth-race-snug-otter.md)
  • Improved /setup-vertex//setup-bedrock to show actual settings.json path and offer a “with 1M context” option
  • Ctrl+U now clears the entire input buffer (previously: delete to start of line); Ctrl+Y to restore

v2.1.110 - April 15, 2026

/tui Flicker-Free Rendering & Push Notifications

  • Added /tui command and tui setting — run /tui fullscreen to switch to flicker-free rendering in the same conversation
  • Added push notification tool — Claude can send mobile push notifications when Remote Control and “Push when Claude decides” are enabled
  • Changed Ctrl+O to toggle normal/verbose transcript only; focus view is now toggled separately via /focus
  • Added autoScrollEnabled config to disable auto-scroll in fullscreen
  • --resume/--continue now resurrect unexpired scheduled tasks
  • /context, /exit, /reload-plugins now work from Remote Control clients
  • Write tool informs the model when you edit proposed content in the IDE diff before accepting
  • Session recap now enabled for users with telemetry disabled (Bedrock, Vertex, Foundry, DISABLE_TELEMETRY)
  • Hardened “Open in editor” actions against command injection from untrusted filenames
  • Fixed MCP tool calls hanging indefinitely when the server connection drops mid-response on SSE/HTTP
  • Fixed non-streaming fallback retries causing multi-minute hangs when the API is unreachable

v2.1.109 - April 15, 2026

Extended-Thinking Progress Hint

  • Improved the extended-thinking indicator with a rotating progress hint

v2.1.108 - April 14, 2026

Prompt Caching 1h TTL, Session Recap

  • Added ENABLE_PROMPT_CACHING_1H env var to opt into 1-hour prompt cache TTL on API key, Bedrock, Vertex, and Foundry
  • Added recap feature to provide context when returning to a session, configurable in /config, invocable with /recap
  • Model can discover and invoke built-in slash commands (/init, /review, /security-review) via the Skill tool
  • /undo is now an alias for /rewind
  • /model warns before mid-conversation switches (next response re-reads full history uncached)
  • /resume picker defaults to current-directory sessions; Ctrl+A shows all projects
  • Improved error messages: server rate limits distinguished from plan limits; 5xx/529 errors link to status.claude.com
  • Reduced memory footprint for file reads/edits/syntax highlighting by loading language grammars on demand

v2.1.107 - April 14, 2026

Thinking Hint Polish

  • Show thinking hints sooner during long operations

v2.1.105 - April 13, 2026

EnterWorktree path, PreCompact Block, Plugin Monitors

  • Added path parameter to the EnterWorktree tool to switch into an existing worktree of the current repository
  • Added PreCompact hook support: hooks can now block compaction by exiting with code 2 or returning {"decision":"block"}
  • Added background monitor support for plugins via a top-level monitors manifest key that auto-arms at session start or on skill invoke
  • /proactive is now an alias for /loop
  • Improved stalled API stream handling: streams abort after 5 minutes of no data and retry non-streaming instead of hanging
  • Improved file write display: long single-line writes (e.g. minified JSON) truncated in UI
  • Improved /doctor layout with status icons; press f to have Claude fix reported issues
  • Improved skill description handling: listing cap raised from 250 to 1,536 characters
  • Improved WebFetch to strip <style>/<script> so CSS-heavy pages don’t exhaust budget before reaching text
  • Fixed images attached to queued messages being dropped
  • Fixed inbound channel notifications being silently dropped after the first message
  • Fixed stdio MCP emitting malformed (non-JSON) output hanging the session instead of failing fast

v2.1.101 - April 10, 2026

/team-onboarding, OS CA Trust, Rate-Limit Polish

  • Added /team-onboarding command to generate a teammate ramp-up guide from local usage
  • OS CA certificate store trust by default — enterprise TLS proxies work without extra setup (opt out via CLAUDE_CODE_CERT_STORE=bundled)
  • /ultraplan and other remote-session features auto-create a default cloud environment instead of requiring web setup first
  • Improved brief mode to retry once when Claude responds with plain text instead of a structured message
  • Improved focus mode: Claude writes more self-contained summaries since it knows you see only its final message
  • Improved rate-limit retry messages to show which limit was hit and when it resets
  • Fixed command injection vulnerability in the POSIX which fallback used by LSP binary detection
  • Fixed permissions.deny rules not overriding a PreToolUse hook’s permissionDecision: "ask"
  • Fixed hardcoded 5-minute request timeout aborting slow backends regardless of API_TIMEOUT_MS
  • Fixed claude -w <name> failing with “already exists” after stale worktree cleanup
  • Fixed subagents in isolated worktrees being denied Read/Edit access to files inside their own worktree
  • Fixed RemoteTrigger tool’s run action sending an empty body

v2.1.100 - April 10, 2026

Paste Improvements, Plugin Dependency Resolution

  • Faster paste handling and better CRLF handling in long pastes
  • Plugin dependency auto-install during marketplace add
  • Performance fixes across SDK subprocess cleanup

v2.1.98 - April 9, 2026

Google Vertex Setup Wizard, Perforce Mode, Monitor Tool

  • Added interactive Google Vertex AI setup wizard from the login screen — guides through GCP auth, project/region config, credential verification, model pinning
  • Added CLAUDE_CODE_PERFORCE_MODE env var — Edit/Write/NotebookEdit fail on read-only files with a p4 edit hint instead of silently overwriting
  • Added Monitor tool for streaming events from background scripts
  • Added subprocess sandboxing with PID namespace isolation on Linux via CLAUDE_CODE_SUBPROCESS_ENV_SCRUB, plus CLAUDE_CODE_SCRIPT_CAPS to cap per-session script invocations
  • Added --exclude-dynamic-system-prompt-sections flag to print mode for cross-user cache improvement
  • Added workspace.git_worktree to the status line JSON when inside a linked git worktree
  • Added W3C TRACEPARENT env var to Bash tool subprocesses when OTEL tracing is enabled
  • Fixed a Bash tool permission bypass where backslash-escaped flags could be auto-allowed as read-only leading to arbitrary code execution
  • Fixed compound Bash commands bypassing forced permission prompts in auto/bypass-permissions modes
  • Fixed --dangerously-skip-permissions being silently downgraded to accept-edits mode
  • Fixed managed-settings allow rules remaining active after admin removed them until restart
  • Improved /agents with tabbed layout: Running tab + Library tab with Run agent / View running instance

v2.1.97 - April 8, 2026

NO_FLICKER Focus View, Cedar Highlighting

  • Added focus view toggle (Ctrl+O) in NO_FLICKER mode showing prompt, one-line tool summary with edit diffstats, and final response
  • Added refreshInterval status line setting to re-run the command every N seconds
  • Added ● N running indicator in /agents next to agent types with live subagent instances
  • Added syntax highlighting for Cedar policy files (.cedar, .cedarpolicy)
  • Fixed MCP HTTP/SSE connections accumulating ~50 MB/hr of unreleased buffers on reconnect
  • Improved image handling: pasted and attached images compressed to same token budget as Read-tool images
  • Improved sandbox: sandbox.network.allowMachLookup now takes effect on macOS
  • Improved Bridge sessions to show local git repo, branch, and cwd on claude.ai session card
  • Fixed CJK/@-mention completion after punctuation; Japanese/Chinese input no longer requires space before / or @

v2.1.96 - April 8, 2026

Bedrock Auth Hotfix

  • Fixed Bedrock requests failing with 403 "Authorization header is missing" when using AWS_BEARER_TOKEN_BEDROCK or CLAUDE_CODE_SKIP_BEDROCK_AUTH (regression in 2.1.94)

v2.1.94 - April 7, 2026

Mantle-backed Bedrock, Higher Default Effort

  • Added support for Amazon Bedrock powered by Mantle — set CLAUDE_CODE_USE_MANTLE=1
  • Changed default effort level from medium to high for API-key, Bedrock/Vertex/Foundry, Team, and Enterprise users
  • Added compact Slacked #channel header with clickable channel link for Slack MCP send-message calls
  • Added keep-coding-instructions frontmatter for plugin output styles
  • Added hookSpecificOutput.sessionTitle to UserPromptSubmit hooks
  • Plugin skills declared via "skills": ["./"] now use the skill’s frontmatter name
  • Fixed agents stuck after 429 with long Retry-After
  • Fixed Console login on macOS silently failing with “Not logged in” when keychain is locked
  • Improved --resume to resume sessions from other worktrees of the same repo directly

v2.1.92 - April 4, 2026

forceRemoteSettingsRefresh, Bedrock Setup Wizard

  • Added forceRemoteSettingsRefresh policy setting — CLI blocks startup until remote managed settings are freshly fetched (fail-closed)
  • Added interactive Bedrock setup wizard from the login screen’s “3rd-party platform” option
  • Added per-model and cache-hit breakdown to /cost for subscription users
  • /release-notes is now an interactive version picker
  • Remote Control session names use your hostname as the default prefix (e.g. myhost-graceful-unicorn)
  • Removed /tag command
  • Removed /vim command (toggle vim mode via /config → Editor mode)
  • Improved Write tool diff computation speed for large files (60% faster on files with tabs/&/$)
  • Linux sandbox now ships the apply-seccomp helper in both npm and native builds
  • Fixed subagent spawning permanently failing with “Could not determine pane count” after tmux window churn

v2.1.91 - April 2, 2026

MCP Result Size Override, Plugin bin/

  • Added MCP tool result persistence override via _meta["anthropic/maxResultSizeChars"] (up to 500K)
  • Added disableSkillShellExecution setting to disable inline shell execution in skills and slash commands
  • Plugins can now ship executables under bin/ and invoke them as bare commands from the Bash tool
  • Added support for multi-line prompts in claude-cli://open?q= deep links (encoded newlines %0A no longer rejected)
  • Fixed transcript chain breaks on --resume that could lose history when async transcript writes fail silently
  • Fixed plan mode in remote sessions losing track of the plan file after a container restart
  • Improved /claude-api skill guidance for agent design patterns (tool surface, context mgmt, caching)
  • Edit tool now uses shorter old_string anchors, reducing output tokens

v2.1.90 - April 1, 2026

/powerup, PowerShell Hardening

  • Added /powerup — interactive lessons teaching Claude Code features with animated demos
  • Added CLAUDE_CODE_PLUGIN_KEEP_MARKETPLACE_ON_FAILURE to keep marketplace cache when git pull fails (offline env)
  • Added .husky to protected directories (acceptEdits mode)
  • Fixed infinite loop where rate-limit options dialog auto-opened, eventually crashing the session
  • Fixed --resume causing a full prompt-cache miss on first request (regression since v2.1.69)
  • Fixed Edit/Write failing with “File content has changed” when a PostToolUse format-on-save hook rewrites the file
  • Fixed auto mode not respecting explicit user boundaries (“don’t push”, “wait for X before Y”)
  • Hardened PowerShell tool permission checks: trailing & background job bypass, -ErrorAction Break hang, archive-extraction TOCTOU, parse-fail deny-rule degradation
  • SSE transport now handles large streamed frames in linear time (was quadratic)

v2.1.89 - April 1, 2026

Deferred Permission, NO_FLICKER Opt-in, PermissionDenied Hook

  • Added "defer" permission decision to PreToolUse hooks — headless sessions pause at a tool call, resume with -p --resume for hook re-evaluation
  • Added CLAUDE_CODE_NO_FLICKER=1 env var to opt into flicker-free alt-screen rendering with virtualized scrollback
  • Added PermissionDenied hook that fires after auto mode classifier denials — return {retry: true} to tell the model it can retry
  • Added named subagents to @ mention typeahead suggestions
  • Added MCP_CONNECTION_NONBLOCKING=true for -p mode to skip MCP connection wait entirely
  • Auto mode: denied commands show a notification and appear in /permissions → Recent, retry with r
  • Documented TaskCreated hook event and its blocking behavior
  • Changed Edit to work on files viewed via Bash sed -n/cat without requiring a separate Read call first
  • Changed thinking summaries to no longer be generated by default in interactive sessions — set showThinkingSummaries: true to restore
  • Changed hook output over 50K characters to be saved to disk with file path + preview instead of injected into context
  • Fixed autocompact thrash loop — detects context refill after three consecutive compacts and stops with actionable error
  • Fixed nested CLAUDE.md being re-injected dozens of times in long sessions that read many files
  • /buddy is here for April 1st — hatch a small creature that watches you code

v2.1.87 - March 29, 2026

Cowork Dispatch Delivery Fix

  • Fixed messages in Cowork Dispatch not getting delivered

March 2026 Updates

Section titled “March 2026 Updates”

v2.1.85-86 - March 27-28, 2026

Conditional Hooks & Stability

  • Conditional hooks: Added if field for hooks using permission rule syntax (e.g., Bash(git *)) to filter when they run, reducing process spawning overhead
  • MCP OAuth RFC 9728: MCP OAuth now follows Protected Resource Metadata discovery to find the authorization server
  • Added CLAUDE_CODE_MCP_SERVER_NAME and CLAUDE_CODE_MCP_SERVER_URL env vars to MCP headersHelper scripts
  • Added .jj and .sl to VCS directory exclusion lists so Grep and file autocomplete skip Jujutsu/Sapling metadata
  • MEMORY.md index now truncates at 25KB as well as 200 lines
  • Read tool uses compact line-number format and deduplicates unchanged re-reads, reducing token usage
  • Plugins blocked by organization policy can no longer be installed or enabled
  • Fixed --resume failing with orphan tool_result blocks on sessions created before v2.1.85
  • Fixed /compact failing when the conversation has grown too large for the compact request itself
  • VSCode: Fixed extension defaulting Max plan users to Sonnet after OAuth token refresh

v2.1.83-84 - March 25-26, 2026

Transcript Search, Plugin Options & PowerShell

  • Transcript search: Press / in transcript mode (Ctrl+O) to search, n/N to step through matches
  • managed-settings.d/ drop-in directory: Separate teams can deploy independent policy fragments that merge alphabetically
  • Plugin options: manifest.userConfig allows plugins to prompt for configuration at enable time, with sensitive: true values stored in keychain
  • PowerShell tool for Windows as an opt-in preview
  • Added CwdChanged and FileChanged hook events for reactive environment management
  • Added TaskCreated hook that fires when a task is created via TaskCreate
  • Added sandbox.failIfUnavailable setting to exit with error when sandbox cannot start
  • Pasted images now insert [Image #N] chip at cursor for positional reference
  • Rules and skills paths: frontmatter now accepts a YAML list of globs
  • MCP tool descriptions and server instructions capped at 2KB to prevent context bloat
  • Token counts ≥1M now display as “1.5m” instead of “1512.6k”
  • Added idle-return prompt nudging users after 75+ minutes to /clear
  • Agents can declare initialPrompt in frontmatter to auto-submit a first turn
  • VSCode: Added rate limit warning banner with usage percentage and reset time

v2.1.81 - March 22, 2026

Bare Mode & Channels

  • --bare flag for scripted -p calls — skips hooks, LSP, plugin sync, and skill directory walks; requires ANTHROPIC_API_KEY; auto-memory fully disabled
  • --channels permission relay — channel servers that declare the permission capability can forward tool approval prompts to your phone
  • Resuming a session that was in a worktree now switches back to that worktree
  • MCP read/search tool calls collapse into a single Queried <server> line (expand with Ctrl+O)
  • Improved plugin freshness — ref-tracked plugins now re-clone on every load
  • Fixed multiple concurrent sessions requiring repeated re-authentication on OAuth token refresh
  • Fixed voice mode silently swallowing retry failures
  • Disabled line-by-line response streaming on Windows due to rendering issues

v2.1.80 - March 21, 2026

Rate Limits in Statusline & Channels Preview

  • rate_limits in statusline: Display Claude.ai rate limit usage (5-hour and 7-day windows) with used_percentage and resets_at
  • --channels research preview: Allow MCP servers to push messages into your session
  • effort frontmatter support for skills and slash commands to override model effort level
  • Added source: 'settings' plugin marketplace source — declare plugin entries inline in settings.json
  • Fixed --resume dropping parallel tool results — sessions with parallel tool calls now restore all pairs
  • Fixed voice mode WebSocket failures caused by Cloudflare bot detection
  • Improved /effort to show what auto currently resolves to
  • Reduced memory usage on startup in large repositories (~80 MB saved on 250k-file repos)

v2.1.79 - March 20, 2026

Remote Control in VSCode

  • VSCode: /remote-control — bridge your session to claude.ai/code to continue from a browser or phone
  • --console flag for claude auth login for Anthropic Console (API billing) authentication
  • VSCode: Session tabs now get AI-generated titles based on your first message
  • Added “Show turn duration” toggle to /config menu
  • Fixed claude -p hanging when spawned as subprocess without explicit stdin
  • Fixed enterprise users unable to retry on rate limit (429) errors
  • Improved startup memory usage by ~18MB across all scenarios

v2.1.78 - March 19, 2026

PowerShell Tool & Plugin Data Persistence

  • PowerShell tool for Windows as an opt-in preview
  • StopFailure hook that fires when the turn ends due to an API error
  • ${CLAUDE_PLUGIN_DATA} variable for plugin persistent state that survives plugin updates
  • Line-by-line response streaming — response text now streams as it’s generated
  • Added effort, maxTurns, and disallowedTools frontmatter support for plugin-shipped agents
  • Terminal notifications now reach the outer terminal when running inside tmux
  • SendMessage now auto-resumes stopped agents in the background
  • Renamed /fork to /branch (/fork still works as alias)
  • Fixed git log HEAD failing inside sandboxed Bash on Linux
  • Security: Fixed silent sandbox disable when dependencies are missing

v2.1.77 - March 18, 2026

64K Output Tokens & /copy N

  • Increased output token limits: 64k default and 128k upper bound for Claude Opus 4.6 and Sonnet 4.6
  • /copy N copies the Nth-latest assistant response
  • Added allowRead sandbox filesystem setting to re-allow read access within denyRead regions
  • Fixed “Always Allow” on compound bash commands saving a single rule instead of per-subcommand
  • Fixed auto-updater starting overlapping binary downloads, accumulating tens of gigabytes of memory
  • Fixed PreToolUse hooks returning "allow" bypassing deny permission rules including enterprise managed settings
  • Improved --resume on fork-heavy sessions — up to 45% faster loading and ~100-150MB less peak memory
  • Background bash tasks are now killed if output exceeds 5GB

v2.1.76 - March 17, 2026

MCP Elicitation & /effort Command

  • MCP elicitation support: MCP servers can now request structured input mid-task via an interactive dialog (form fields or browser URL)
  • /effort slash command to set model effort level
  • worktree.sparsePaths setting for large monorepos to check out only needed directories via git sparse-checkout
  • Added PostCompact hook that fires after compaction completes
  • Added -n / --name CLI flag to set a display name for the session at startup
  • Fixed deferred tools losing input schemas after conversation compaction
  • Fixed auto-compaction retrying indefinitely — circuit breaker now stops after 3 attempts
  • Improved background agent behavior — killing preserves partial results in context

v2.1.49-50 - March 10-11, 2026

Task Management & ToolSearch Deferred Loading

  • Task management tools: TaskCreate, TaskOutput, TaskDelete for workflow automation
  • CronCreate tool to schedule future tasks via cron syntax
  • ToolSearch deferred loading: All MCP and plugin tools defer loading until model requests them, reducing startup latency
  • Added deniedMcpServers managed setting for enterprise admins to block specific MCP servers
  • Added maxOutputTokens frontmatter support for skills and slash commands
  • Added PostToolUse hook that fires after a tool call completes
  • Fixed MCP tools with no arguments showing spurious {} input indicator
  • Improved /export to include session metadata in the transcript file

v2.1.47-48 - March 8-9, 2026

Quick-Fix Commands & PreCompact Hook

  • Added CLAUDE_CODE_CUSTOM_QUICK_FIX_COMMANDS env var to register custom quick-fix providers
  • Added CLAUDE_CODE_BUILTIN_LSP_TIMEOUT_MS env var to configure LSP diagnostics timeout
  • Added PreCompact hook that fires before conversation compaction begins
  • Added contentType parameter to PluginCall for sending non-JSON request bodies
  • Fixed Grep tool hanging on pathologically large files (>1GB)
  • Fixed Edit tool emitting mangled diff when target file was modified between preview and save
  • Improved plan mode to show estimated execution time and token usage

v2.1.45-46 - March 6-7, 2026

SendMessage Tool & /mcp Command

  • SendMessage tool to send messages to other agents in the conversation
  • /mcp command for interactive MCP server connection management
  • Added ANTHROPIC_DEFAULT_MODEL env var to set a preferred model for the session
  • Added PostStop hook that fires after the turn ends
  • Added sandbox.filesystem.allowWrite support in claude.yaml rules frontmatter
  • Fixed Bash tool hang on macOS when stdout is piped and the process terminates
  • Fixed voice mode freezing the UI for 2+ seconds when recording stops
  • Improved @-mention autocomplete to show recently-edited files first

v2.1.44 - March 5, 2026

Ultra-Thinking Mode & /sandbox Command

  • Ultra-thinking (ultratlk) mode for extended reasoning — opt in with /ultratlk or set ultraThinkingEnabled: true
  • /sandbox command for viewing and toggling sandbox filesystem rules
  • Added ANTHROPIC_EXTENDED_THINKING_BUDGET_TOKENS env var to configure thinking token budget
  • Added maxThinkingTokens frontmatter support for skills and agents
  • Added requiredMcpServers managed setting for enterprise admins to enforce MCP server availability
  • Fixed multi-line user input showing only the first line in the transcript
  • Fixed MCP tools showing “Connection refused” errors when the server initially fails to start
  • Improved MCP server error messages to distinguish between connection failures and protocol errors

v2.1.41-43 - March 3-4, 2026

Plugin Tools & Agent Access Rules

  • Added PluginCall tool for plugins to invoke other plugin capabilities
  • Added allow and deny frontmatter support for skills and agents to configure tool access
  • Added CLAUDE_CODE_ENABLE_EXPERIMENTAL_FEATURES env var to opt into preview features
  • Added MCP resource templateVariables support for dynamic resource discovery
  • Added PostToolUse and PostToolUseFailure hooks for handling tool results and errors
  • Added model frontmatter support for agents to pin a specific model variant
  • Fixed MCP tools with long descriptions causing token budget overflow
  • Improved permission rule matching to handle complex bash syntax (pipes, redirects, substitutions)

v2.1.38-40 - March 1-2, 2026

Worktrees & MCP Enhancements

  • --worktree flag for working in git worktrees
  • Added MCP server health check on startup
  • Added support for conditional permission rules via if field
  • Added ANTHROPIC_API_KEY_HELPER env var for dynamic API key injection
  • Added support for MCP servers with custom TLS certificates
  • Fixed Edit tool creating corrupted diffs on binary files
  • Improved Grep tool output formatting
  • Reduced startup latency by parallelizing plugin loads

February 2026 Updates

Section titled “February 2026 Updates”

v2.1.36-37 - February 20-21, 2026

Fast Mode for Opus 4.6

  • v2.1.37: Fixed an issue where /fast was not immediately available after enabling /extra-usage
  • v2.1.36: Fast mode is now available for Opus 4.6 — same model, faster output. Toggle with /fast

v2.1.34 - February 17, 2026

Agent Teams & Sandbox Fixes

  • Fixed a crash when agent teams setting changed between renders
  • Fixed a bug where commands excluded from sandboxing (via sandbox.excludedCommands or dangerouslyDisableSandbox) could bypass the Bash ask permission rule when autoAllowBashIfSandboxed was enabled

v2.1.33 - February 15, 2026

Agent Hooks & VSCode Remote Sessions

  • Added TeammateIdle and TaskCompleted hook events for multi-agent workflows
  • Added memory frontmatter field for agents — persistent memory with user, project, or local scope
  • Added support for restricting sub-agents via Task(agent_type) syntax in agent “tools” frontmatter
  • Added plugin name to skill descriptions and /skills menu for better discoverability
  • VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
  • VSCode: Added git branch and message count to the session picker, with search by branch name
  • Fixed multiple issues: thinking interruption, API proxy 404 fallback, proxy env var propagation, /resume XML markup display

v2.1.32 - February 12, 2026

Claude Opus 4.6 & Agent Teams

  • Claude Opus 4.6 is now available!
  • Agent Teams research preview — multi-agent collaboration (set CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
  • Automatic Memories — Claude now records and recalls memories as it works
  • Summarize from here — partial conversation summarization via message selector
  • Skills from .claude/skills/ in additional directories (--add-dir) now loaded automatically
  • Skill character budget scales with context window (2% of context)
  • Fixed @ file completion paths in subdirectories, heredoc template literal errors, Thai/Lao spacing vowels

v2.1.31 - February 8, 2026

Session Resume & Stability Fixes

  • Added session resume hint on exit, showing how to continue your conversation later
  • Fixed PDF too large errors permanently locking up sessions
  • Fixed bash commands incorrectly reporting “Read-only file system” errors with sandbox mode enabled
  • Fixed temperatureOverride being silently ignored in the streaming API path
  • Improved system prompts to guide model toward using dedicated tools instead of bash equivalents
  • Reduced layout jitter when the spinner appears/disappears during streaming

v2.1.30 - February 3, 2026

PDF Page Ranges & Token Metrics

  • PDF Page Ranges: Specify exact page ranges when reading PDFs (e.g., @file.pdf:1-5)
  • OAuth Improvements: Enhanced OAuth flow with better error handling and refresh token management
  • Token Metrics: New /usage displays detailed token metrics including input/output breakdown
  • MCP Server Health: Improved health checks and automatic restart for failing MCP servers
  • Bug Fixes:
    • Fixed session state corruption in multi-agent workflows
    • Fixed terminal rendering issues with wide characters in Asian languages
    • Fixed memory leak in long-running background agents

v2.1.27-29 - January 28-31, 2026

PR Workflow Enhancements

  • v2.1.29: --from-pr flag to start Claude Code with context from a GitHub PR
  • v2.1.28: Automatic PR linking - Claude Code suggests related PRs and issues
  • v2.1.27: Enhanced context management - smarter auto-compaction preserves critical context
  • Bug Fixes: Fixed race condition in concurrent agent spawning, improved Windows path handling

v2.1.25 - January 25, 2026

Arrow Key History & Sandbox Improvements

  • Arrow Key History: Use up/down arrows to navigate through prompt history
  • /sandbox Improvements: Better isolation, persistent state across sessions, improved cleanup
  • Wide Character Fixes: Complete fix for CJK character display and input
  • Performance: 40% faster file indexing for large repositories

v2.1.20-24 - January 20-24, 2026

Various Improvements

  • v2.1.24: Custom agent avatars and display names
  • v2.1.23: Improved /diff command with side-by-side view
  • v2.1.22: Better error messages with suggested fixes
  • v2.1.21: Enhanced web search with source attribution
  • v2.1.20: MCP server hot-reload without restart

January 2026 Updates

Section titled “January 2026 Updates”

v2.1.19 - January 19, 2026

SKILL.md Support

  • SKILL.md: Define custom skills in .claude/skills/ directory with SKILL.md manifest files
  • Skill Discovery: /skills command lists available skills with descriptions
  • Skill Parameters: Define inputs, outputs, and required context for each skill
  • Documentation: Skills can include examples and usage patterns

v2.1.17-18 - January 16-18, 2026

Session Forking & Processor Compatibility

  • v2.1.18: Session forking - create branches from any point in conversation history
  • v2.1.17: Improved processor compatibility for ARM64 and Apple Silicon M3/M4
  • Cloud Handoff: New & suffix to hand off tasks to cloud agents
  • Plan Mode: Enhanced /plan with dependency tracking and parallel execution hints
  • Bug Fixes: Fixed agent timeout handling, improved OAuth callback reliability

Version 2.0.x Series

Section titled “Version 2.0.x Series”

December 2025 Updates

Section titled “December 2025 Updates”

v2.0.67 - December 11, 2025

AI-Powered Prompt Suggestions

  • Prompt Suggestions: Claude now suggests prompts to speed up your workflow - press Tab to accept or Enter to submit
  • Thinking Mode Default: Thinking mode is now enabled by default for Opus 4.5
  • Thinking Config: Thinking mode configuration has moved to /config
  • Permissions Search: Added search functionality to /permissions command with / keyboard shortcut
  • Diagnostics: Show reason why autoupdater is disabled in /doctor
  • Bug Fixes:
    • Fixed false “Another process is currently updating Claude” error
    • Fixed MCP servers from .mcp.json stuck in pending state in non-interactive mode
    • Fixed word deletion/navigation with non-Latin text (Cyrillic, Greek, Arabic, Hebrew, Thai, Chinese)
    • Windows: Fixed plugin MCP servers failing due to colons in log directory paths

v2.0.65 - December 9, 2025

Model Switching While Typing

  • Quick Model Switch: Switch models while writing a prompt using alt+p (Linux/Windows) or option+p (macOS)
  • Status Line: Added context window information to status line input
  • File Suggestions: Added fileSuggestion setting for custom @ file search commands
  • Shell Override: Added CLAUDE_CODE_SHELL environment variable to override automatic shell detection
  • Bug Fixes: Fixed prompt not being saved to history when aborting with Escape

v2.0.64 - December 7, 2025

Named Sessions & Rules Directory

  • Auto-Compacting: Made auto-compacting instant
  • Async Operations: Agents and bash commands can run asynchronously and send messages to wake up the main agent
  • Stats Command: /stats now provides interesting stats: favorite model, usage graph, usage streak
  • Named Sessions:
    • Use /rename to name sessions
    • Use /resume <name> in REPL or claude --resume <name> from terminal
  • Rules Directory: Added support for .claude/rules/ directory (see memory documentation)
  • Image Metadata: Added image dimension metadata when images are resized
  • Resume Screen: Improved /resume screen with grouped forked sessions and keyboard shortcuts (P for preview, R for rename)
  • VS Code: Added copy-to-clipboard button on code blocks and bash tool inputs
  • Bedrock: Improved efficiency of token counting, added support for aws login credentials

v2.0.62 - December 5, 2025

Attribution Settings

  • Recommended Indicator: Added “(Recommended)” indicator for multiple-choice questions
  • Attribution Setting: Added attribution setting to customize commit and PR bylines (deprecates includeCoAuthoredBy)
  • Bug Fixes: Fixed duplicate slash commands, skill files symlink issues, lock file stale issues

v2.0.60 - December 3, 2025

Background Agent Support

  • Background Agents: Agents now run in background while you work
  • CLI Flag: Added --disable-slash-commands flag
  • Commit Attribution: Added model name to “Co-Authored-By” commit messages
  • MCP Toggle: Enabled /mcp enable [server-name] or /mcp disable [server-name] to quickly toggle servers
  • Fetch Optimization: Updated Fetch to skip summarization for pre-approved websites
  • VS Code: Added support for multiple terminal clients connecting to IDE server simultaneously

November 2025 Updates

Section titled “November 2025 Updates”

v2.0.59 - November 28, 2025

Agent Configuration

  • Agent CLI Flag: Added --agent CLI flag to override agent setting for current session
  • Agent Setting: Added agent setting to configure main thread with specific agent’s system prompt, tool restrictions, and model
  • VS Code Fix: Fixed .claude.json config file being read from incorrect location

v2.0.58 - November 26, 2025

Opus 4.5 for Pro Users

  • Pro Access: Pro users now have access to Opus 4.5 as part of their subscription!
  • Timer Fix: Fixed timer duration showing “11m 60s” instead of “12m 0s”
  • Windows: Managed settings now prefer C:\Program Files\ClaudeCode if it exists

v2.0.57 - November 24, 2025

Plan Feedback & Streaming

  • Plan Feedback: Added feedback input when rejecting plans
  • VS Code: Added streaming message support for real-time response display

v2.0.56 - November 22, 2025

Terminal Progress & VS Code Sidebar

  • Progress Bar: Added setting to enable/disable terminal progress bar (OSC 9;4)
  • VS Code: Added support for VS Code’s secondary sidebar (VS Code 1.97+)

v2.0.55 - November 20, 2025

Improved UX

  • Proxy Fix: Fixed proxy DNS resolution being forced on by default (now opt-in via CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true)
  • Auto-Submit: Improved AskUserQuestion tool to auto-submit single-select questions on the last question
  • Fuzzy Matching: Improved fuzzy matching for @ file suggestions

v2.0.51 - November 14, 2025

Claude Opus 4.5 & Desktop App

  • Opus 4.5: Added Claude Opus 4.5! (Announcement)
  • Desktop App: Introducing Claude Code for Desktop (Download)
  • Plan Mode: Plan Mode now builds more precise plans and executes more thoroughly
  • Usage Notifications: Usage limit notifications easier to understand
  • Usage Display: Switched /usage back to ”% used”

Version 1.0.x Series (Production)

Section titled “Version 1.0.x Series (Production)”

August 2025 Updates

Section titled “August 2025 Updates”

v1.0.70 - August 6, 2025

Performance and Windows Fixes

  • Performance: Optimized message rendering for better performance with large contexts
  • Windows: Fixed native file search, ripgrep, and subagent functionality
  • @-mentions: Added support for @-mentions in slash command arguments

v1.0.69 - August 5, 2025

Model Upgrade

  • Opus 4.1: Upgraded Opus to version 4.1 with enhanced capabilities

v1.0.68 - August 4, 2025

Multiple Improvements

  • Model Names: Fixed incorrect model names being used for commands like /pr-comments
  • Windows Enhancements:
    • Improved permissions checks for allow/deny tools and project trust
    • Better sub-process spawning to eliminate “No such file or directory” errors with pnpm
  • Diagnostics: Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
  • SDK: Added canUseTool callback support for tool confirmation
  • Settings: Added disableAllHooks setting
  • Performance: Improved file suggestions performance in large repos

v1.0.65 - July 31, 2025

IDE and Windows Stability

  • IDE: Fixed connection stability issues and error handling for diagnostics
  • Windows: Fixed shell environment setup for users without .bashrc files

v1.0.64 - July 30, 2025

Agents and SDK Updates

  • Agents: Added model customization support - specify which model an agent should use
  • Agents: Fixed unintended access to the recursive agent tool
  • Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
  • SDK: Fixed user input tracking across multi-turn conversations
  • Search: Added hidden files to file search and @-mention suggestions

v1.0.63 - July 28, 2025

Windows Functionality Fixes

  • Fixed file search, @agent mentions, and custom slash commands functionality on Windows

v1.0.62 - July 26, 2025

Agent Mentions and Hooks

  • Agent Support: Added @-mention support with typeahead for custom agents (@<your-custom-agent>)
  • Hooks: Added SessionStart hook for new session initialization
  • Commands: /add-dir command now supports typeahead for directory paths
  • Network: Improved network connectivity check reliability

v1.0.61 - July 24, 2025

Transcript Mode and Settings

  • Transcript Mode: Changed Esc to exit transcript mode rather than interrupt (Ctrl+R)
  • Settings: Added --settings flag to load settings from a JSON file

v1.0.60 - July 23, 2025

Custom Subagents

  • Major Feature: Create custom subagents for specialized tasks! Run /agents to get started

v1.0.59 - July 22, 2025

SDK and Hooks Enhancements

  • SDK: Added tool confirmation support with canUseTool callback
  • SDK: Allow specifying env for spawned process
  • Hooks: Exposed PermissionDecision to hooks (including “ask”)
  • Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
  • Model Fix: Fixed issue where some Max users specifying Opus would still see fallback to Sonnet

v1.0.58 - July 20, 2025

PDF Support and Improvements

  • PDF Reading: Added support for reading PDF files
  • MCP: Improved server health status display in ‘claude mcp list’
  • Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands

v1.0.57 - July 19, 2025

Model Selection and Permissions

  • Slash Commands: Added support for specifying a model in slash commands
  • Permissions: Improved permission messages to help Claude understand allowed tools
  • Fix: Remove trailing newlines from bash output in terminal wrapping

v1.0.56 - July 18, 2025

Windows and WSL Improvements

  • Windows: Enabled Shift+Tab for mode switching on Node.js versions supporting terminal VT mode
  • WSL: Fixes for WSL IDE detection
  • AWS: Fixed issue causing awsRefreshHelper changes to .aws directory not to be picked up

v1.0.55 - July 17, 2025

Model Clarifications and Fixes

  • Models: Clarified knowledge cutoff for Opus 4.5 and Sonnet 4.5 models
  • Windows: Fixed Ctrl+Z crash
  • SDK: Added ability to capture error logging
  • System Prompt: Added --system-prompt-file option to override system prompt in print mode

July 2025 Updates

Section titled “July 2025 Updates”

v1.0.54 - July 16, 2025

Major Improvements

  • Hooks Enhancement: Added UserPromptSubmit hook with current working directory in hook inputs
  • Slash Commands: Support for argument-hint field in custom command frontmatter
  • Windows Improvements:
    • OAuth flow now uses port 45454 with correct browser URL construction
    • Mode-switching shortcut changed to Alt + M (fixing Shift+Tab conflict)
    • Plan mode renders properly in Windows terminals
  • Shell Execution: In-memory shell snapshot prevents file-related errors

v1.0.53 - July 11, 2025

Game-Changing Update

  • Context Expansion: @-mention file content limit increased from 100 to 2000 lines
  • AWS Integration: Helper scripts for credential refresh:
    • awsAuthRefresh for SSO login operations
    • awsCredentialExport for background STS operations

v1.0.52 - July 10, 2025

  • MCP Instructions: Support for Model Context Protocol server instructions
  • Continued MCP integration improvements

v1.0.51 - July 8, 2025

Native Windows Support Launch

  • Windows Native: Runs without WSL (Git for Windows required for Git features)
  • Cloud Integration: Amazon Bedrock support via AWS_BEARER_TOKEN_BEDROCK
  • Diagnostics: /doctor command identifies and fixes invalid configurations
  • Interactive Mode: --append-system-prompt flag now works in interactive mode
  • Performance: Auto-compaction threshold increased from 60% to 80%
  • Telemetry: Enhanced OpenTelemetry metadata (OS, architecture, WSL version)
  • Bug Fixes:
    • Directory names with spaces handled correctly
    • Custom commands in subdirectories register with proper namespacing
    • Plan mode rejection edge case resolved

v1.0.48 - July 7, 2025

Stability and UX Improvements

  • Critical Fix: Resolved freeze-on-launch bug from v1.0.45
  • Bash Tool UX: Real-time progress showing last 5 lines of output
  • MCP Config: Environment variable expansion in server configs
  • Shell Reliability: Snapshots moved to ~/.claude directory
  • New Hook: PreCompact hook for pre-compaction actions
  • Vim Mode: Extended keybindings (c, f/F, t/T motions)

v1.0.44-45 - July 3, 2025

Search and Export Features

  • v1.0.44:
    • New /export command for conversation transcripts
    • MCP resource_link results display properly
    • Ctrl+Z now suspends (resume with fg)
    • Prompt undo remapped to Ctrl+U
  • v1.0.45:
    • Search Tool Overhaul: Redesigned grep tool with new parameters
    • Jupyter notebook diff view disabled (prevents timeouts)
    • Atomic config writes prevent corruption
    • Hooks refinement: Stop vs SubagentStop events
    • Custom command namespacing restored

June 2025 Updates

Section titled “June 2025 Updates”

v1.0.38 - June 30, 2025

Hooks System Introduction

  • User-Defined Hooks: Inject custom behavior at lifecycle events
  • Enables logging, approvals, and integrations
  • Community-driven feature based on GitHub feedback
  • Major extensibility expansion

v1.0.27 - June 18, 2025

Real-Time MCP Integration

  • SSE Support: Server-Sent Events for MCP servers
  • HTTP Transport: Additional transport option
  • OAuth 2.0: Authentication for remote MCP servers
  • Resource Syntax: @server:protocol://path for external data
  • Live connections to external tool servers

v1.0.24-25 - June 25, 2025

Reliability Improvements

  • v1.0.24:
    • Enhanced /mcp command formatting
    • Fixed array merging in nested settings
  • v1.0.25:
    • Fixed: Slash commands not appearing/executing
    • Custom command discovery overhaul
    • Ghostty proxy compatibility
    • Web search reliability improvements

v1.0.23 - June 11, 2025

CLI Package Update

  • Official NPM Package:
    • CLI Tool: @anthropic-ai/claude-code (NPM)
    • Note: No official Python package currently available
  • Enhanced CLI integration features
  • Improved streaming output support

v1.0.17-18 - June 9, 2025

Multi-Directory Support

  • v1.0.17:
    • Sub-task messages in non-interactive mode
    • parent_tool_use_id for task correlation
    • VS Code diff tool crash fixes
    • Process title shows as claude
  • v1.0.18:
    • --add-dir flag for additional directories
    • Streaming stdin support
    • CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable
    • SSE auto-reconnection for MCP servers

v1.0.10-11 - June 4, 2025

General Availability Features

  • Access Expansion: Available to Pro and Max plans
  • v1.0.10: Markdown table rendering support
  • v1.0.11:
    • /upgrade command for plan switching
    • Enhanced authentication UI
    • Improved to-do list management

May 2025 Updates

Section titled “May 2025 Updates”

v1.0.0 - May 22, 2025

General Availability Launch

  • Status: Graduated from preview to supported product
  • Models: Claude Opus 4.5 and Sonnet 4.5 integration
  • Coincided with Claude 3.7 “Sonnet” model launch
  • v1.0.1:
    • DISABLE_INTERLEAVED_THINKING flag option
    • Provider-specific model names in UI

Version 0.2.x Series (Beta/Preview)

Section titled “Version 0.2.x Series (Beta/Preview)”

May 2025 Beta Updates

Section titled “May 2025 Beta Updates”

v0.2.125 - May 21, 2025

Final Beta - Breaking Changes

  • Bedrock ARN: Changed from %2F to literal /
  • Logging: DEBUG=true replaced with ANTHROPIC_LOG=debug
  • Preparation for GA release

v0.2.117 - May 20, 2025

  • JSON Output: Changed to nested message objects (breaking change)
  • Settings: cleanupPeriodDays configuration added
  • Debug Mode: Global --debug flag introduced

v0.2.108 - May 18, 2025

Interactive Thinking

  • Real-Time Steering: Send messages while Claude is “thinking”
  • Bash Timeouts: BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS
  • Fixed thinking mode in non-interactive mode
  • MCP wizard deprecated for dedicated commands

v0.2.105 - May 16, 2025

Web Browsing Capability

  • Web Search: New fetch tool for URLs
  • Status Command: /status replaces welcome message info
  • Vim Motions: Word-level navigation (w, b, e)
  • Performance optimizations for startup and file operations

v0.2.102 - May 14, 2025

  • Thinking Mode: Improved reliability and triggers
  • @-mentions: Better handling of images and directories
  • Input: Multiple large text chunks can be pasted

v0.2.93 - May 1, 2025

Session Management

  • Continue Sessions: claude --continue for recent session
  • Resume by ID: claude --resume <ID> for specific sessions
  • To-Do List: Internal task management for complex operations

April 2025 Beta Updates

Section titled “April 2025 Beta Updates”

v0.2.82 - April 25, 2025

  • Permissions: --disallowedTools flag and config
  • Tool Renaming: Clearer names (LSTool→LS, View→Read)

v0.2.75 - April 20, 2025

Major Interactivity Update

  • Concurrent Queries: Queue messages while Claude is busy
  • Image Support: Drag-and-drop or paste images
  • @-mentions: Reference files directly in prompts
  • MCP One-off: --mcp-config <path> flag
  • Autocomplete: Performance improvements for large projects

v0.2.70 - April 15, 2025

  • Network Tools: curl and similar commands unlocked
  • Parallel Search: Multiple web queries simultaneously
  • Interrupt: ESC key immediately stops Claude in auto-accept mode

March 2025 Beta Updates

Section titled “March 2025 Beta Updates”

v0.2.54 - March 25, 2025

  • Memory: # prefix for persistent context
  • Long Output: Ctrl+R reveals full tool output
  • MCP SSE: Server-Sent Events transport support

v0.2.47 - March 15, 2025

Auto-Compaction & QoL

  • Autocomplete: Tab completion for filenames
  • Auto-Accept: Shift+Tab toggles file edit approvals
  • Infinite Sessions: Automatic conversation compaction
  • Configurable via /config interface

v0.2.44 - March 14, 2025

Thinking Mode Introduction

  • Deep Reasoning: Triggered by “think” in prompts
  • Step-by-step planning before execution
  • “Think harder” and “ultrathink” variations

v0.2.34 - March 8, 2025

Vim Mode & Custom Commands

  • Vim Keybindings: Toggle with /vim
  • Custom Commands: .claude/commands/ directory support
  • Normal and insert modes with hjkl navigation

February 2025

Section titled “February 2025”

v0.2.0 - February 24, 2025

Research Preview Launch

  • Initial Release: Agentic coding tool for terminal
  • Core capabilities:
    • Codebase scanning and understanding
    • Intelligent code search
    • File editing with permissions
    • Test suite execution
    • Git operations (commit, push, branch)
    • Shell command execution
  • Permission system for user control
  • Internal testing: 45-minute tasks completed in one pass

Version Patterns & Release Cycle

Section titled “Version Patterns & Release Cycle”

Release Frequency

  • Preview Phase: Multiple releases per week
  • Post-GA: 2-3 releases per week average
  • Hotfixes: Within 24-48 hours when needed

Version Numbering

  • Major: Significant feature additions
  • Minor: New capabilities and improvements
  • Patch: Bug fixes and small enhancements

Breaking Changes

  • Clearly marked in changelog
  • Migration guides provided
  • Rare after GA (v1.0.0)

Notable Feature Introductions Timeline

Section titled “Notable Feature Introductions Timeline”
  1. February 2025: Core CLI capabilities and permission system
  2. March 2025: Vim mode, custom commands, thinking mode, auto-compaction
  3. April 2025: Images, @-mentions, concurrent queries, network tools
  4. May 2025: Web browsing, session management, GA launch with new models
  5. June 2025: SDKs, Hooks system, real-time MCP, multi-directory support
  6. July 2025: Native Windows, 2000-line context, enhanced AWS integration, custom subagents
  7. August 2025: Opus 4.1 upgrade, performance optimizations, enhanced diagnostics
  8. November 2025: Claude Opus 4.5, Claude Code for Desktop, Pro users get Opus 4.5
  9. December 2025: Background agents, named sessions, .claude/rules/, prompt suggestions, model switching (alt+p)
  10. January 2026: SKILL.md support, session forking, cloud handoff (&), --from-pr flag, arrow key history
  11. February 2026: Claude Opus 4.6, agent teams research preview, fast mode, automatic memories, PDF page ranges

Staying Updated

Section titled “Staying Updated”
Check for Updates Run 'claude update' to get the latest version
View Release Notes Use '/release-notes' command within Claude Code
GitHub Releases Follow detailed releases on GitHub
Privacy Policy Terms of Service