AI & LLM Vulnerabilities

Coder's Azure identity endpoint was vulnerable to SSRF (server-side request forgery, where an attacker tricks a server into making requests to unintended targets) because it accepted unsigned certificates and fetched arbitrary URLs without validation. An attacker could craft a fake certificate pointing to any internal or external address, forcing the Coder server to connect to it and reveal whether the target was reachable through error messages, enabling network reconnaissance and potential attacks on internal services.

A domain allowlist (list of approved websites) in the Apify Model Context Protocol server is bypassed because it uses simple string prefix matching instead of proper URL validation. An attacker can create a fake subdomain like `https://docs.apify.com.evil.com/` that passes the check, allowing the tool to fetch arbitrary content from attacker-controlled servers and return it to the AI, which can lead to prompt injection (tricking the AI by hiding instructions in fetched content) and potential account compromise.

Envoy AI Gateway has a vulnerability where it improperly parses JSON-RPC messages (a protocol for remote procedure calls) in a case-insensitive way, even though the specification requires case-sensitive matching. This allows attackers to send messages with duplicate fields using different capitalization (like 'name' and 'Name'), causing the gateway to alter and forward a different request than what was originally sent, potentially bypassing security checks in systems that use this gateway.

A security flaw in n8n (a workflow automation tool) allowed authenticated users to bypass restrictions on which websites could receive sensitive credentials, potentially exposing them. The vulnerability was in an endpoint (a URL that accepts requests) that didn't properly check the intended security rules before sending data to external servers.

NiceGUI has a vulnerability in two routes (resource and ESM module routes) that serve files without authentication. If a request tries to access a directory instead of a file through these routes, it causes an unhandled error that writes a large traceback (around 100 lines) to the server log. An attacker can repeatedly trigger this to fill up disk space, overload logging systems, and create false alarms in monitoring without needing any special access.

OpenTelemetry eBPF Instrumentation has a memory leak in its `CappedConcurrentHashMap` class, which is used to track Java TLS connections. When entries are deleted from the map, the keys are not removed from an internal queue, causing the queue to grow indefinitely in systems with many short-lived connections. This can eventually cause the Java process to run out of memory.

Claude HUD version 0.0.12 and earlier has a vulnerability where it creates terminal hyperlinks (clickable links in terminal windows) using user-controlled data without properly cleaning it first, allowing attackers to inject malicious terminal codes (ANSI codes, which control formatting and behavior in terminals) that could change text colors, fake command prompts, steal clipboard data, or redirect users to attacker-controlled websites.

OpenTelemetry eBPF Instrumentation (OBI) has a memory leak vulnerability where a CPU mismatch causes the system to use a 256-byte fallback buffer but still tries to read up to 8KB of data from it, reading beyond the buffer's boundaries and leaking adjacent memory into telemetry (data about system performance). This happens in the HTTP tracing path when context propagation is enabled and certain conditions are met.