Banks Prompt.chat_messages() JSON Parsing Flaw

Publication date: 2026-10-08

Last updated on: 2026-10-08

Assigner: GitHub, Inc.

Description

Banks generates meaningful LLM prompts using a simple template language. Prior to 2.5.0, Banks Prompt.chat_messages() attempts to parse every line of rendered template output as ChatMessage JSON. When an application renders untrusted data and passes the returned ChatMessage objects to an LLM provider, attacker-controlled JSON can cross the prompt boundary and become a system, assistant, or tool message because ChatMessage.role accepts arbitrary strings. This can override application instructions, alter the intended prompt structure, or confuse downstream tool and message handling. This issue is fixed in version 2.5.0.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published

2026-10-08

Last Modified

2026-10-08

Generated

2026-10-09

AI Q&A

2026-10-09

EPSS Evaluated

N/A

NVD

CVE-2026-107717

Affected Vendors & Products

Vendor	Product	Version / Range
masci	banks	< 2.5.0

Helpful Resources

Exploitability

CWE

KEV

CWE ID	Description
CWE-20	The product receives input or data, but it does not validate or incorrectly validates that the input has the properties that are required to process the data safely and correctly.

Attack-Flow Graph

Executive Summary

Banks is a tool that generates LLM prompts using templates. Before version 2.5.0, its Prompt.chat_messages() function incorrectly parsed every line of rendered template output as ChatMessage JSON. This allowed attacker-controlled JSON to be treated as system, assistant, or tool messages, potentially altering prompt structure or instructions.

Impact Analysis

If exploited, this vulnerability could allow attackers to inject malicious JSON into prompts, overriding application instructions or confusing downstream tool handling. This might lead to unintended LLM behavior, data leaks, or incorrect outputs depending on how the application uses Banks.

Mitigation Strategies

Upgrade Banks to version 2.5.0 or later to address the vulnerability in Prompt.chat_messages(). Avoid passing untrusted data through ChatMessage JSON parsing to prevent attacker-controlled input from altering prompt structure or instructions.

Hi! I’m here to help you understand CVE-2026-107717. Ask me anything about the vulnerability, its impact, or mitigation strategies.

0/70