← Incident Database
Tool Abuse / Excessive AgencyCritical
LangChain LLMMathChain arbitrary code execution (CVE-2023-29374)
April 2023 · LangChain
What happened
LLMMathChain passed LLM-generated text into Python exec/eval to evaluate math. A crafted prompt could make the LLM emit Python that escaped the math context and executed arbitrary code on the host (CVSS 9.8).
Root cause
Insecure use of exec/eval on LLM output with no sandboxing, a classic prompt-injection-to-code-execution.
Fix / outcome
Fixed in LangChain 0.0.142 and later with safer numeric evaluation.
Sources
Learn this attack class
This incident is an example of Tool Abuse / Excessive Agency. Read the guide, then try it hands-on in the Academy.