Claude Code v2.1.69 Released - New Features and Security Fixes
Claude Code v2.1.69 brings new features including /claude-api skill, expanded Voice STT language support, and numerous security and bug fixes.
Claude Code v2.1.69 Released - New Features and Security Fixes
Claude Code v2.1.69 was released on March 5, 2026 (US time). This release includes numerous new features along with security fixes and bug corrections.
Key Changes
New Commands and Skills
The /claude-api skill was added to facilitate building applications with the Claude API and Anthropic SDK. The /reload-plugins command was also introduced, allowing plugin changes to be applied without restarting Claude Code.
Voice STT Language Expansion
Voice STT (speech-to-text) now supports 10 new languages, bringing the total to 20. Added languages include Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, and Norwegian.
Remote Control Enhancement
The /remote-control command now accepts an optional name argument to set a custom session title visible in claude.ai/code.
UI/UX Improvements
Effort level display (e.g., “with low effort”) was added to the logo and spinner, making it easier to identify the current effort setting. Agent name now displays in the terminal title when using —agent.
Security Fixes
A security vulnerability where nested skill discovery could load skills from gitignored directories like node_modules was fixed. The trust dialog issue where all .mcp.json servers were silently enabled on first run was also resolved.
Other New Features
- Added sandbox.enableWeakerNetworkIsolation setting (macOS only) to allow Go programs like gh, gcloud, and terraform to verify TLS certificates when using a custom MITM proxy
- Added ${CLAUDE_SKILL_DIR} variable for skills to reference their own directory in SKILL.md content
- Added InstructionsLoaded hook event and agent_id, agent_type to hook events
Technical Details
- File count cache now properly handles glob ignore patterns
- Fixed —setting-sources user not blocking dynamically discovered project skills
- Multiple memory leak fixes (REPL render scopes, React Compiler memoCache, interactive mode hook events)
関連記事
Claude Code v2.1.64 リリース — Remote Control新機能とパフォーマンス改善
Claude Code v2.1.64がリリース。Remote Controlサブコマンド追加、BashToolパフォーマンス改善、コンテキスト自動保存機能強化など多数の改善を含む。
Claude Code v2.1.66 リリース — 新规スラッシュコマンドとHTTP Hooksの導入
Claude Code v2.1.66がリリース。/simplify・/batchスラッシュコマンドの追加、HTTP Hooksサポート、多数のパフォーマンス改善とメモリリーク修正を含む。
Claude Code v2.1.68 リリース — Opus 4.6デフォルト設定とultrathinkキーワードの再導入
Claude Code v2.1.68がリリース。Opus 4.6がMax/Teamサブスクライバーにmedium effortでデフォルト設定され、ultrathinkキーワードが再導入。Opus 4と4.1は削除。
人気記事
Cursor Automations リリース - トリガーとスケジュールに基づく常駐エージェント
Cursorが新しいAutomations機能をリリース。トリガーとスケジュールに基づいて動作する常駐エージェントを構築できる。
Cursor v2.7 リリース - JetBrains IDEでのCursor利用が可能に
Cursor v2.7では、Agent Client Protocol (ACP)を通じてIntelliJ IDEA、PyCharm、WebStormなどのJetBrains IDEでCursorが利用可能になった。
Cursor v2.6 リリース - MCP Appsとチーム向けプラグイン管理機能
Cursor v2.6では、エージェントチャット内でのインタラクティブなUI、プライベートプラグインのチーム共有機能強化、デバッグモードの改善などが発表された。
最新記事
Claude Code v2.1.72 リリース - 自動メモリ保存とループ機能追加
AnthropicがClaude Code v2.1.72をリリース。自動メモリ保存、/copyコマンド、/loopコマンド、Cronスケジュールツール、Voiceキーバインディングなど多数の新機能を搭載。
Gemini 3.1 Pro — コーディング・推論性能で業界トップを争夺
GoogleのGemini 3.1 Proは、2026年2月に発表された最新推論モデル。SWE-Bench Verified 80.6%、LiveCodeBenchで2887スコアを達成し、Claude Opus 4.6と並ぶトップレンジモデルとして注目される。
OpenAI Symphony — エージェントの監視から「仕事の管理」へのパラダイムシフト
OpenAIが新プロジェクトSymphonyを公開。GitHub Issue (Linear)から自動的にエージェントを実行し、PRを作成まで行う自律型オートメーションツール。