Support for Homomorphic Encryption to Secure Credential Handling

Currently there is a security limitation—passwords stored in Credential Configuration nodes cannot be securely passed to Python Script nodes, as scripts could inadvertently expose or misuse plaintext credentials. This restriction makes perfect sense from a security perspective.

However, I’d like to suggest exploring the possibility of integrating homomorphic encryption (HE) into KNIME to address this challenge.

Why Homomorphic Encryption?

Homomorphic encryption allows computations directly on encrypted data without requiring decryption, thus maintaining strict confidentiality of sensitive information throughout the workflow.

Example Use Case:

  • Credentials are encrypted using homomorphic encryption methods in KNIME.
  • Encrypted credentials can safely pass into Python Script nodes.
  • Python scripts perform encrypted computations (like validation or checks) without ever accessing plaintext data.
  • Results remain encrypted until securely decrypted elsewhere within the KNIME environment, avoiding any exposure of sensitive credentials.

Benefits:

  • Enhanced security and compliance
  • Allows more flexible and secure integration with Python scripts
  • Maintains strict protection of sensitive credentials throughout workflows

Integrating HE libraries (e.g., TenSEAL, Microsoft SEAL) might be a suitable technical direction.

Thanks for considering this suggestion!

HE’ing the password makes no sense because you wont be able to access it.

if your data is protected with HE you can utilize Python script or Java Snippet given you have the libraries present.

also: a lot easier solutions e.g. include

  • loading the decryption key as part of the python node
  • don’t use python but build the logic in Knime / Java