Vulnerability · Published Jul 9, 2026 · Assigned by GitHub_M
CVE-2026-54769Langroid: Sandbox Escape to Remote Code Execution via Incomplete `eval()` Mitigation in TableChatAgent
Critical 10.0
In plain English
A critical-severity code injection in langroid from langroid. Proof-of-concept exploit code is public. The chance it is exploited in the next 30 days is estimated at 0.9% (higher than 60% of all known vulnerabilities). Fixed versions are available - upgrade the affected package.
Official description
Langroid is a framework for building large-language-model-powered applications. Versions prior to 0.65.2 are vulnerable to a critical Sandbox Escape leading to Remote Code Execution (RCE) in its `TableChatAgent` and `VectorStore` capabilities. When these agents evaluate LLM-generated tool messages with `full_eval=True`, they attempt to sandbox the execution by explicitly setting `locals` to an empty dictionary `{}` inside Python's `eval()` function. However, this relies on an incomplete understanding of Python's execution model. Because `__builtins__` is not explicitly scrubbed from the `globals` dictionary mapping, Python implicitly injects all built-ins during execution, granting full access to functions like `__import__('os').system()`. Since `TableChatAgent.pandas_eval()` executes external LLM outputs natively, this bypass permits any attacker providing prompt payload to achieve unauthenticated RCE on the host system. Version 0.65.2 patches the issue.
Affected packages & fixed versions
| Ecosystem | Package | Fixed in | Advisory |
|---|---|---|---|
| PyPI | langroid | 0.65.2 | GHSA-q9p7-wqxg-mrhc |
| PyPI | langroid | 0.65.2 | PYSEC-2026-2581 |
Affected products (as reported)
Weakness: CWE-94 · Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
References
- https://github.com/langroid/langroid/security/advisories/GHSA-q9p7-wqxg-mrhc x_refsource_CONFIRM