CriticalVulnerabilityLLM-specific
CVE-2025-46725: Langroid LanceDocChatAgent code execution via pandas eval
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2025-46725
- Published
- Record updated
- Affected
- langroid < 0.53.15
- Fixed in
- 0.53.15
- released May 17, 2025, 3 days before disclosure
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 0.5%
Summary
Langroid, a Python framework for building LLM-powered applications, prior to version 0.53.15 uses pandas eval() through compute_from_docs() in LanceDocChatAgent. An attacker may be able to make the agent run malicious commands via QueryPlan.dataframe_calc, compromising the host system.
Mitigation
Fixed in 0.53.15: the affected function now sanitizes input by default against the most common attack vectors, and the project documentation adds warnings about the risky behavior.
Affected packages in the Exposure Registry
Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.
- langroidPyPILLM dependency since 2023-07-14
Related items
- MediumCVE-2026-121936: easyVoice server-side request forgery through openaiBaseUrl in /api/v1/ttsSimilar attack · NVD/CVE Database
- MediumCVE-2026-118628: MLflow get-artifacts server-side request forgery information disclosureSimilar attack · NVD/CVE Database
- HighCVE-2026-118625: MLflow Tracking Server privilege escalation via get_endpointsSimilar attack · NVD/CVE Database
- InfoWhen trust becomes the payload: Inside the threats, etc quarterly reportSimilar attack · Elastic Security Labs
- MediumCritical Langflow flaws enable authentication bypass and RCESimilar attack · CSO Online