/

CVE-2026-77810

Severity:

CRITICAL RISK

(9.9)

Code Injection (CWE-95) in Neptune Athena Connector

CVE-2026-77810

Awaiting Analysis

Vulnerability type

Category:

Code Injection

CWE:

CWE-95

Impact:

Exposure of Lambda configuration and potentially sensitive execution properties

Severity:

CRITICAL RISK

(9.9)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A critical code injection vulnerability in the Neptune connector may allow a user who can access Neptune through Athena Federated Query to access properties associated with the Lambda function that supplies connector compute.

The issue becomes risky when connector-controlled input is evaluated in a way that exposes the Lambda execution environment or its configuration to an authorized-but-untrusted query user.

Affected context

This vulnerability occurs in deployments that use the Neptune connector through Athena Federated Query and rely on a Lambda function to provide connector compute. The risk depends on the connector evaluating attacker-influenced data or query-related content within the Lambda execution context, potentially making Lambda properties accessible to users who otherwise only have Neptune query access.

Exploit preconditions

  • Attack surface: Neptune access through Athena Federated Query
  • Required access: An account or principal authorized to use the affected Neptune connector path
  • Trigger point: Processing connector input or query-related data in the supplying Lambda function
  • Why exploitation is possible: Improper evaluation of user-influenced content may expose Lambda properties beyond the intended Neptune data-access boundary

Why this happens

The underlying issue is consistent with CWE-95: Improper Neutralization of Directives in Dynamically Evaluated Code ('Eval Injection'). The connector can allow input associated with an authorized query workflow to influence dynamic code evaluation or equivalent execution behavior within its Lambda runtime. This breaks the intended separation between query data and executable directives, potentially exposing runtime properties to the invoking user.

Potential impact

  • Users may access properties associated with the connector's Lambda function
  • Sensitive configuration or environment details could be disclosed
  • Credentials or service metadata may be exposed if present in the Lambda context
  • An attacker could use disclosed information to support further attacks against connected AWS resources
  • The intended isolation between Neptune access and connector-runtime access could be weakened

General mitigation guidance

Mitigation generally focuses on removing dynamic evaluation of user-influenced content and preserving a strict boundary between query data and executable logic.

Common defensive patterns include:

  • Replacing evaluation-based processing with explicit parsers and allowlisted operations
  • Treating query values, identifiers, and connector parameters as data rather than code
  • Restricting Lambda environment properties and avoiding storage of unnecessary secrets in runtime configuration
  • Applying least-privilege IAM permissions to the connector Lambda role
  • Separating data-query permissions from access to operational or configuration metadata
  • Using a vendor-provided remediated release and validating the connector behavior in a controlled environment

How teams detect issues like this

Organizations can look for this issue through:

  • Static analysis rules identifying eval-like APIs, dynamic expression execution, and construction of executable code from query-controlled values
  • Code review checks for connector parameters or database content reaching dynamic execution paths
  • Runtime monitoring for unexpected access to Lambda environment properties, configuration APIs, or metadata
  • IAM and policy analysis that identifies connector roles with broader permissions than required
  • Automated security review tools that flag dynamic evaluation and sensitive runtime-property access

Sources

SourceLink
aws.amazon.comaws.amazon.com/security/security-bulletins/2026-087-aws/
github.comgithub.com/awslabs/aws-athena-query-federation/releases/tag/v2026.30.1
github.comgithub.com/awslabs/aws-athena-query-federation/security/advisories/GHSA-v7c2-5wfg-qg44

Frequently asked

questions

It may be exploitable through the Athena Federated Query access path by a user who can use the affected Neptune connector. Direct network access to the Lambda function is not necessarily required.

The attacker generally needs authorization to access Neptune through Athena Federated Query and to reach the vulnerable connector workflow.

Eliminate dynamic evaluation of user-influenced content, use explicit allowlisted parsing and operations, and deploy a vendor-provided remediated connector release.

Keep query data separate from executable logic, minimize Lambda permissions and stored secrets, review eval-like APIs, and monitor access to runtime configuration and metadata.

Find out what's already
exploitable in your codebase.