« Back to list

CVE-2026-25879

Status: DeferredCritical (9.8)—

Langroid is a framework for building large-language-model-powered applications. Prior to version 0.63.0, SQLChatAgent executes SQL produced by an LLM, which is influenceable by prompt injection. When configured with a database role that has privileges enabling code execution or filesystem access (e.g., PostgreSQL pg_execute_server_program, MySQL FILE, MSSQL xp_cmdshell), an attacker who can shape the agent's input — including indirectly via data returned to the LLM — can coerce execution of dialect-specific primitives such as `COPY ... FROM PROGRAM`, achieving RCE on the database host.

Read full descriptionShow less

Fixed in v0.63.0 by defaulting SQLChatAgent to a SELECT-only sqlglot-parsed statement allowlist with a dialect-aware dangerous-pattern blocklist; allow_dangerous_operations=True restores the previous unrestricted behavior for trusted deployments.

CVSS

Exploitation probability (EPSS)

EPSS (Exploit Prediction Scoring System, FIRST) estimates how likely a vulnerability is to be exploited in the wild within 30 days. It complements CVSS (impact) and CISA KEV (confirmed exploitation).

🎯 ATT&CK techniques

How this vulnerability is exploited and what the attacker gains, in MITRE ATT&CK terms.

SQLChatAgent ejecuta SQL generado por LLM vulnerable a inyección de prompts sin sanitizar (CWE-89, CWE-94); acceso red sin autenticación (AV:N/PR:N) permite RCE en host de BD mediante primitivas de ejecución de código SQL (pg_execute_server_program, xp_cmdshell).

Inferred by our analysis agent from the official description, CVSS vector and CWE, and checked by a supervisor. May contain errors.

🛡️ ATT&CK mitigations that cover these techniques

Privileged Account ManagementVulnerability ScanningRestrict Web-Based ContentRestrict File and Directory PermissionsRemote Data StorageNetwork Segmentation

Affected technologies (1)

⚠ AI-inferred from the description — NVD hasn't analyzed this CVE yet, these aren't verified CPEs.

Langroid — LangroidAI

CWEs

References

Raw JSON (NVD)

Show
{
  "id": "CVE-2026-25879",
  "cveTags": [],
  "metrics": {
    "ssvcV203": [
      {
        "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
        "ssvcData": {
          "id": "CVE-2026-25879",
          "role": "CISA Coordinator",
          "options": [
            {
              "exploitation": "poc"
            },
            {
              "automatable": "yes"
            },
            {
              "technicalImpact": "total"
            }
          ],
          "version": "2.0.3",
          "timestamp": "2026-06-02T15:03:18.990011Z"
        }
      }
    ],
    "cvssMetricV31": [
      {
        "type": "Secondary",
        "source": "security-advisories@github.com",
        "cvssData": {
          "scope": "UNCHANGED",
          "version": "3.1",
          "baseScore": 9.8,
          "attackVector": "NETWORK",
          "baseSeverity": "CRITICAL",
          "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
          "integrityImpact": "HIGH",
          "userInteraction": "NONE",
          "attackComplexity": "LOW",
          "availabilityImpact": "HIGH",
          "privilegesRequired": "NONE",
          "confidentialityImpact": "HIGH"
        },
        "impactScore": 5.9,
        "exploitabilityScore": 3.9
      }
    ]
  },
  "affected": [
    {
      "source": "security-advisories@github.com",
      "affectedData": [
        {
          "vendor": "langroid",
          "product": "langroid",
          "versions": [
            {
              "status": "affected",
              "version": "< 0.63.0"
            }
          ]
        }
      ]
    }
  ],
  "published": "2026-06-01T23:16:21.930",
  "references": [
    {
      "url": "https://github.com/langroid/langroid/security/advisories/GHSA-mxfr-6hcw-j9rq",
      "source": "security-advisories@github.com"
    },
    {
      "url": "https://github.com/langroid/langroid/security/advisories/GHSA-mxfr-6hcw-j9rq",
      "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0"
    }
  ],
  "vulnStatus": "Deferred",
  "weaknesses": [
    {
      "type": "Secondary",
      "source": "security-advisories@github.com",
      "description": [
        {
          "lang": "en",
          "value": "CWE-89"
        },
        {
          "lang": "en",
          "value": "CWE-94"
        }
      ]
    }
  ],
  "descriptions": [
    {
      "lang": "en",
      "value": "Langroid is a framework for building large-language-model-powered applications. Prior to version 0.63.0, SQLChatAgent executes SQL produced by an LLM, which is influenceable by prompt injection. When configured with a database role that has privileges enabling code execution or filesystem access (e.g., PostgreSQL pg_execute_server_program, MySQL FILE, MSSQL xp_cmdshell), an attacker who can shape the agent's input — including indirectly via data returned to the LLM — can coerce execution of dialect-specific primitives such as `COPY ... FROM PROGRAM`, achieving RCE on the database host. Fixed in v0.63.0 by defaulting SQLChatAgent to a SELECT-only sqlglot-parsed statement allowlist with a dialect-aware dangerous-pattern blocklist; allow_dangerous_operations=True restores the previous unrestricted behavior for trusted deployments."
    },
    {
      "lang": "es",
      "value": "Langroid es un framework para construir aplicaciones impulsadas por modelos de lenguaje grandes. Antes de la versión 0.63.0, SQLChatAgent ejecuta SQL producido por un LLM, el cual es influenciable mediante inyección de prompts. Cuando se configura con un rol de base de datos que tiene privilegios que permiten la ejecución de código o el acceso al sistema de archivos (por ejemplo, PostgreSQL pg_execute_server_program, MySQL FILE, MSSQL xp_cmdshell), un atacante que puede moldear la entrada del agente - incluso indirectamente a través de datos devueltos al LLM - puede forzar la ejecución de primitivas específicas del dialecto como 'COPY ... FROM PROGRAM', logrando RCE en el host de la base de datos. Corregido en la v0.63.0 al establecer por defecto SQLChatAgent a una lista de permitidos de sentencias analizadas por sqlglot de solo SELECT con una lista de bloqueados de patrones peligrosos consciente del dialecto; allow_dangerous_operations=True restaura el comportamiento irrestricto anterior para implementaciones de confianza."
    }
  ],
  "lastModified": "2026-07-21T19:10:00.107",
  "sourceIdentifier": "security-advisories@github.com"
}