AIKIDO-2026-768825
banks is vulnerable to Improper Input Validation
65
Medium Risk
This Affects:
banksTL;DR
Prompt.chat_messages() renders a template and then parses every output line as JSON, so template data that matches the ChatMessage schema is turned into a chat message. User controlled values added into a prompt can therefore declare their own system or assistant role, forge tool definitions, or inject content blocks that the caller forwards to the LLM as trusted messages. This lets untrusted input override instructions and cross prompt trust boundaries. The fix marks output produced by the chat tag and trusted filters and rejects any unmarked line.
Who does this affect?
You are affected if you are using a version that falls within the vulnerable range and you render untrusted data into prompt templates that are converted to chat messages.
Background info
banks is vulnerable to Improper Input Validation in versions 1.0.0 - 2.4.5.
How to fix this
Upgrade the banks library to the patch version.
Are You Affected?
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.
I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001CompliantISO 42001Compliant
FedRAMPImplementing© All Intel data is openly available and commercially licensed.