Unofficial AI-summarized news site (not affiliated with any AI company)
AI News JP / www.ai-news.jp
🟠 Important AI Summary · Source: Claude Code Releases

6 security fixes bundled—Bash bypass, sandbox escape, permission prompt hiding. Prod users should update.

Claude Code v2.1.223: Security Fixes & Permission Management Improvements

Original: v2.1.223

Importance: セキュリティ脆弱性6件と権限管理の複数不具合が同時修正され、本番環境ユーザーに直接影響する。

Summary

Anthropic released Claude Code v2.1.223 with 17 changes, including 6 security patches addressing Bash permission bypass, workflow sandbox escapes, and permission dialog circumvention. New features: owner wildcard entries for marketplace management, `/teleport` hint for local continuation, `modelOverrides` robustness, context auto-compaction improvements. Also fixes session resumption glitches, Linux sandbox startup failures, and gateway model discovery issues.

Key Points

  • 6 security fixes: Bash bypass, workflow sandbox escape, permission prompt hiding
  • Owner wildcard entries for marketplace allow/block at GitHub org level
  • All 1M-window models now auto-compact to 200K; unknown model IDs stay in bounds
  • Bundled fixes for session resume, Linux sandbox startup, gateway model discovery
  • `/teleport` command enables continuation of cloud session locally with session ID
View developer notes (APIs, breaking changes, migration)

v2.1.223 fixes 6 security issues: Bash permission bypass via padding/Unicode, workflow `import()` sandbox escape, `bypassPermissions` ignoring org disable policy, permission prompt info hiding, forked agent resumption failures, malformed diagnostics hangs. Features: `owner/*` wildcard in `strictKnownMarketplaces`/`blockedMarketplaces`, `/teleport <session_id>` for local continuation, `CLAUDE_CODE_DISABLE_1M_CONTEXT` now affects all 1M-native models (200K auto-compact), `modelOverrides` treats unknown keys as ignored. Linux sandbox startup, model discovery, and context compression robustness enhanced.

Claude Code安全性/研究API/SDKAudience: 開発者Audience: 企業導入担当

Source: https://github.com/anthropics/claude-code/releases/tag/v2.1.223

Outlet: Claude Code Releases

This article is an AI-generated summary (OpenAI GPT-4o-mini) of publicly available information from Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, Sakana, and other vendors. The original source URL is always provided in accordance with fair-use citation requirements. Summaries are AI-generated and may contain mistranslations or misinterpretations. Always verify details with the original source.