Claude Code version 2.1.288 exhibited a security vulnerability where it failed to block destructive commands like `rm -rf build` when sandbox rules were set to deny or ask. This allowed the command to execute successfully in all tested instances. In contrast, version 2.1.289, released shortly after, successfully blocked these commands, indicating a fix for the vulnerability. The issue stemmed from the sandbox not correctly interpreting commands prefixed with environment variables or those preceded by bare variable assignments. AI
IMPACT Fixes a critical security flaw in Claude Code's sandbox, enhancing the safety of AI-assisted coding environments.
RANK_REASON This is a bug fix for a specific tool, not a new frontier release or significant industry event.
Read on dev.to — Claude Code tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →