AI Sec Watch

Directus, a content management system, failed to properly sanitize sensitive data (like user tokens, two-factor authentication secrets, and API keys) before storing them in revision history records. This meant that anyone with access to the revision database table could read these secrets in plaintext, potentially allowing account takeover or unauthorized access to third-party services.

Directus has a security flaw in its TUS resumable upload endpoint (a feature that lets users upload files in chunks) that lets any authenticated user overwrite any file in the system by specifying its UUID (unique identifier), bypassing row-level permissions (rules like 'users can only edit their own files'). This can lead to permanent data loss and allow low-privilege users to replace important files with malicious content.

The mobile_open_url tool in mobile-mcp doesn't check what type of URL scheme (the protocol prefix like http:// or tel://) it receives before sending it to Android, allowing attackers to use prompt injection (tricking an AI by hiding instructions in its input) to execute dangerous commands like making phone calls, sending SMS messages, or accessing private data on a connected mobile device.

Anthropic is changing its policy so Claude users can no longer use their subscription to access OpenClaw (a third-party tool that integrates with Claude), forcing them to pay separately instead. The change takes effect April 4th, and may be motivated by Anthropic wanting to promote its own competing tools like Claude Cowork.

BentoML's Dockerfile generation uses an unsandboxed Jinja2 template engine (a tool that processes template files with dynamic code) with dangerous extensions enabled, allowing attackers to embed malicious code in a template file. When a victim imports a malicious bento archive and runs the containerize command, the attacker's code executes directly on the victim's host machine before any container isolation happens, rather than inside a container where it would be restricted.

BentoML has a command injection vulnerability in its cloud deployment setup script where user-supplied system packages are inserted directly into shell commands without proper escaping. An attacker can craft a malicious bentofile.yaml file that executes arbitrary commands on BentoCloud's build infrastructure (the servers that prepare applications for deployment) when the application is deployed, potentially stealing secrets or compromising the infrastructure.

This research examines how attackers could exploit Amazon Bedrock's multi-agent systems (groups of specialized AI agents working together) through prompt injection (tricking an AI by hiding malicious instructions in user input), potentially discovering agent instructions and executing unauthorized tool actions. The study found no vulnerabilities in Bedrock itself, but highlighted a broader LLM challenge: these systems cannot reliably distinguish between legitimate developer instructions and adversarial user input. The research was conducted ethically on owned systems in collaboration with Amazon's security team.

LiteLLM had a security flaw where JWT authentication (a method to verify user identity using encoded tokens) could be bypassed through a cache key collision. When JWT authentication was enabled, the system only used the first 20 characters of a token as a cache key, and since different tokens from the same signing algorithm could have identical first 20 characters, an attacker could create a fake token matching a legitimate user's cached token and gain their permissions. The flaw only affects deployments with JWT/OIDC authentication explicitly enabled, which is not the default configuration.

LiteLLM had a security flaw where an authenticated user could access a configuration endpoint (`/config/update`) without needing admin permissions, allowing them to modify settings, run malicious code, read files, or take over admin accounts. The vulnerability affected any user who already had login access to the system.