CVE-2026-4231

A vulnerability was found in vanna-ai vanna up to 2.

Verified by Precogs Threat Research
Last Updated: Mar 16, 2026
Base Score
7.3HIGH

Executive Summary

CVE-2026-4231 is a high severity vulnerability affecting appsec, ai-code. It is classified as Server-Side Request Forgery. Ensure your systems and dependencies are patched immediately to mitigate exposure risks.

Precogs AI Insight

"The root cause of this vulnerability lies in within A vulnerability, allowing insufficient sanitization protocols during data parsing. This flaw provides a direct pathway for attackers to intercept or modify sensitive data flows before they reach secure enclaves. Precogs AI Analysis Engine utilizes semantic code analysis to harden the environment against lateral movement."

Exploit Probability (EPSS)
Low (0.1%)
Public POC
Undisclosed
Exploit Probability
Elevated (52%)
Public POC
Available
Affected Assets
appsecai codeCWE-918

What is this vulnerability?

CVE-2026-4231 is categorized as a critical SQL Injection flaw. Based on our vulnerability intelligence, this issue occurs when the application fails to securely handle untrusted data boundaries.

A vulnerability was found in vanna-ai vanna up to 2.0.2. Affected by this vulnerability is the function update_sql/run_sql of the file src/vanna/legacy/fla...

This architectural defect enables adversaries to bypass intended security controls, directly manipulating the application's execution state or data layer. Immediate strategic intervention is required.

Risk Assessment

MetricValue
CVSS Base Score7.3 (HIGH)
Vector StringCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L
PublishedMarch 16, 2026
Last ModifiedMarch 16, 2026
Related CWEsCWE-918

Impact on Systems

Data Exfiltration: Full compromise of the database schema, allowing extraction of all tables, user records, and PII.

Authentication Bypass: Attackers can manipulate boolean logic in authentication queries to log in as administrators.

Remote Code Execution: In severe configurations (e.g., xp_cmdshell in MSSQL), attackers can execute shell commands on the database underlying OS.

How to fix this issue?

Implement the following strategic mitigations immediately to eliminate the attack surface.

1. Prepared Statements Migrate entirely to parameterized queries (Prepared Statements) or an Object-Relational Mapper (ORM) to decouple code from data.

2. Input Validation Implement rigorous allow-list input validation for all sorting, filtering, and query parameters.

3. Principle of Least Privilege Ensure the database service account has the minimum necessary privileges, restricting DROP, TRUNCATE, and system execution commands.

Vulnerability Signature

// Example of a vulnerable Node.js/Express snippet

const category = req.query.category;

// DANGEROUS: Direct string concatenation of user input
const query = `SELECT * FROM products WHERE category = '$\{category\}'`;

db.query(query, (err, result) =\> \{
  if (err) throw err;
  console.log(result);
\});

// SECURED: Using parameterized queries avoids SQL injection
const category = req.query.category; // Ensure scope appropriately

// Safe: The database driver treats '?' strictly as data, not executable code
const query = 'SELECT * FROM products WHERE category = ?';

db.query(query, [category], (err, result) =\> \{
  if (err) throw err;
  console.log(result);
\});

References and Sources

Vulnerability Code Signature

Attack Data Flow

StageDetail
SourceUser-supplied URL parameter
VectorServer fetches the user-controlled URL
SinkHTTP request library (e.g., fetch, axios)
ImpactAccess to internal services, metadata endpoints, or port scanning

Vulnerable Code Pattern

// ❌ VULNERABLE: Server-Side Request Forgery
app.get('/proxy', async (req, res) => {
  const targetUrl = req.query.url;
  // Taint sink: unvalidated outbound request
  const response = await fetch(targetUrl);
  res.send(await response.text());
});

Secure Code Pattern

// ✅ SECURE: Allowlist validation
const ALLOWED_DOMAINS = ['api.partner.com'];
app.get('/proxy', async (req, res) => {
  const targetUrl = new URL(req.query.url);
  
  if (!ALLOWED_DOMAINS.includes(targetUrl.hostname)) {
    return res.status(403).send("Domain not allowed");
  }
  
  // Sanitized outbound request
  const response = await fetch(targetUrl.href);
  res.send(await response.text());
});

How Precogs Detects This

Precogs API Security Engine actively detects SSRF vulnerabilities by tracking unvalidated URL sinks and enforcing strict allowlist policies.\n

Related Vulnerabilitiesvia CWE-918

CVE-2026-3543110 CRITICAL

Server-Side Request Forgery (SSRF) in Server-side request forgery (ssrf) in Microsoft Entra ID Entitlement Management allows an unauthorized attacker to perform spoofing over a network

CWE-918
CVE-2026-322109.3 CRITICAL

Server-Side Request Forgery (SSRF) in Server-side request forgery (ssrf) in Microsoft Dynamics 365 (Online) allows an unauthorized attacker to perform spoofing over a network

CWE-918
CVE-2026-3218610 CRITICAL

Server-Side Request Forgery (SSRF) in Server-side request forgery (ssrf) in Microsoft Bing allows an unauthorized attacker to elevate privileges over a network

CWE-918
CVE-2026-48740 LOW

CVE-2026-4874: Server-Side Request Forgery in Keycloak

CWE-918
CVE-2026-45287.3 HIGH

A vulnerability was determined in trueleaf ApiFlow 0.

CWE-918
CVE-2024-562796.5 MEDIUM

Server-Side Request Forgery (SSRF) — Cloud metadata access

CWE-918

Is your system affected?

Precogs AI detects CVE-2026-4231 in compiled binaries, LLMs, and application layers — even without source code access.