CVE-2026-46590: Apache Camel PQC Unsafe Deserialization Remote Code Execution

Technical summary

CVE-2026-46590 is an unsafe deserialization vulnerability in Apache Camel's camel-pqc component affecting HashicorpVaultKeyLifecycleManager, AwsSecretsManagerKeyLifecycleManager, and FileBasedKeyLifecycleManager. These implementations use raw java.io.ObjectInputStream.readObject() to deserialize Base64-encoded KeyMetadata from backend storage without an ObjectInputFilter or class allow-list. The type cast to KeyMetadata occurs only after deserialization completes, allowing arbitrary readObject() side effects to execute before the type check. The vulnerability is classified as CWE-502 (Deserialization of Untrusted Data) and represents an incomplete remediation of CVE-2026-40048, which addressed file-based storage but did not add proper input filtering or cover the Vault and AWS manager implementations. Affected versions: 4.18.0–4.18.2 and 4.19.0–4.20.x.

Business impact

Compromise of a key management backend storage system becomes a direct path to remote code execution within the cryptographic key management layer of an organization's Camel deployment. An attacker with write access to HashiCorp Vault or AWS Secrets Manager (via a leaked token or IAM credential) can achieve code execution in the context of the application managing post-quantum keys, potentially exposing encrypted data, disrupting key rotation workflows, or pivoting to other systems. For organizations relying on PQC for future-proofing cryptographic infrastructure, this undermines the security model by allowing compromise at the key management tier rather than requiring attacks on the application or transport layer.

Affected systems

Apache Camel versions 4.18.0 through 4.18.2 (4.18.x LTS stream) and 4.19.0 through 4.20.x are vulnerable. The issue is resolved in Apache Camel 4.18.3 (LTS branch) and 4.21.0 (current). Any deployment using the camel-pqc component with HashiCorp Vault, AWS Secrets Manager, or file-based key lifecycle managers is at risk if the backend storage system is accessible to a principal with write permissions.

Exploitability

Exploitability requires an attacker to have authenticated write access to the configured key backend: a valid HashiCorp Vault token with kv put permissions on the PQC metadata path, or AWS IAM credentials with secretsmanager:PutSecretValue action on the relevant secret. The attack is non-interactive—once a crafted serialized object is stored, it is automatically deserialized during normal key-lifecycle operations, triggering code execution without further user action. CVSS score of 8.8 (HIGH) reflects the requirement for authenticated backend access combined with the high impact of remote code execution.

Remediation

Immediate action: upgrade to Apache Camel 4.21.0 (recommended for all branches) or 4.18.3 (if on the 4.18.x LTS stream). For deployments unable to patch immediately, implement strict access controls on the key backend: restrict write permissions via HashiCorp Vault policies (e.g., deny except for the application's own service identity) and AWS IAM policies (limit secretsmanager:PutSecretValue to the key-management application only). Additionally, segregate PQC key material into a backend separate from general secrets to reduce blast radius if less-trusted principals gain write access to shared storage.

Patch guidance

Upgrade Apache Camel to version 4.21.0 for all currently supported deployments. Organizations on the 4.18.x LTS stream should upgrade to 4.18.3. Verify the patch version in your build artifacts or application manifest before deployment. Test the upgrade in a non-production environment to ensure compatibility with your PQC key lifecycle workflows and any custom KeyLifecycleManager implementations. Consult the Apache Camel release notes for any breaking changes between your current version and the patched version.

Detection guidance

Monitor HashiCorp Vault audit logs and AWS CloudTrail for PutSecretValue or kv put operations on paths containing PQC key metadata—especially unexpected writes from service accounts other than the intended key-management application. Watch application logs for deserialization errors, ClassNotFoundException, or unexpected exceptions during key-lifecycle operations, as these may indicate an attempted exploit. Network-level monitoring can flag unusual outbound connections from the key-management process immediately after key retrieval. Review IAM policies and Vault ACLs to confirm that only the appropriate application identity has write access to key storage.

Why prioritize this

This is a HIGH-severity vulnerability with a direct path to remote code execution in a security-critical component (post-quantum cryptographic key management). Although it requires prior authentication to the backend storage system, the risk is amplified by the common practice of sharing credentials across teams and the fact that many organizations may not yet have segregated write access to secrets backends. The incomplete fix for CVE-2026-40048 signals that this issue has not been fully eradicated in prior patches, making prompt upgrade essential to avoid regression.

Risk score, explained

CVSS 8.8 reflects: Network-accessible attack vector (deserialization during key-lifecycle operations), low attack complexity (straightforward serialized-object injection), low privileges required (authenticated access to backend), no user interaction, unrestricted scope within the application context, and high impact across confidentiality, integrity, and availability (arbitrary code execution). The score appropriately captures the severity of remote code execution in a key-management context while acknowledging the prerequisite of backend write access.

Frequently asked questions