AI & LLM Vulnerabilities

Claude Code had a security flaw where it would read settings from a file (`.claude/settings.json`) that could be controlled by someone creating a malicious repository, allowing them to bypass the workspace trust dialog (a security prompt that asks for permission before running code). This meant an attacker could trick users into running code without their knowledge or consent. The vulnerability has been patched.

Cloud CLI (a user interface for Claude Code and similar tools) had a critical vulnerability in versions before 1.25.0 where user inputs called projectPath, initialCommand, and sessionId were directly used to build system commands without filtering, allowing attackers to inject arbitrary OS commands (OS command injection, where an attacker tricks the system into running unauthorized commands) through WebSocket connections. This vulnerability has been patched in version 1.25.0.

Cloud CLI (a user interface for accessing Claude Code and similar tools) has a vulnerability in versions before 1.24.0 where user input in the git configuration endpoint is not properly sanitized before being executed as shell commands. This means an authenticated attacker (someone with login access) could run arbitrary OS commands (commands that do whatever they want on the operating system) by exploiting how backticks, command substitution (${}), and backslashes are interpreted within the double-quoted strings.

OpenClaw, a personal AI assistant, had a security flaw in versions 2026.2.13 and below on macOS where OAuth tokens (authentication credentials that prove you're logged in) could be used to inject malicious OS commands (commands that run at the operating system level) into the credential refresh process. An attacker could exploit this by crafting a specially designed token to execute arbitrary commands on the affected system.

The Claude CLI tool on macOS had a shell injection vulnerability (a security flaw where attackers can run arbitrary commands) in how it stored authentication tokens in the system keychain. The problem occurred because user-controlled OAuth tokens were directly inserted into shell commands without proper protection, allowing an attacker to break out of the intended command and execute malicious code.

sf-mcp-server, a tool that connects Salesforce to Claude for Desktop, has a command injection vulnerability (CWE-78, a flaw where attackers inject malicious commands into user input). The vulnerability exists because the software unsafely uses child_process.exec (a function that runs shell commands) with user-controlled input, allowing attackers to execute arbitrary shell commands with the server's privileges.

Claude Code (an AI tool that can write and modify software) before version 2.1.7 had a security flaw where it could bypass file access restrictions through symbolic links (shortcuts that point to other files). If a user blocked Claude Code from reading a sensitive file like /etc/passwd, the tool could still read it by accessing a symbolic link pointing to that file, bypassing the security controls.

Claude Code is an agentic coding tool (software that can automatically write and execute code) that had a vulnerability in versions before 2.0.72 where attackers could bypass safety confirmation prompts and execute untrusted commands through the find command by injecting malicious content into the tool's context window (the input area where the AI reads information). The vulnerability has a CVSS score (a 0-10 severity rating) of 7.7, meaning it is considered high severity.

Claude Code, a tool that helps AI write and execute code automatically, had a security flaw before version 1.0.111 where it didn't properly check website addresses (URLs) before making requests to them. The app used a simple startsWith() check (looking only at the beginning of a domain name), which meant attackers could register fake domains like modelcontextprotocol.io.example.com that would be mistakenly trusted, allowing the tool to send data to attacker-controlled sites without the user knowing.