In plain English: A security firewall for AI agents. It keeps instructions and data strictly separate, screens prompts for attacks, checks the tools your agents use, and produces a signed compliance record for every request.
Quantum OSCL sits in front of your agents and stops prompt-injection and tool-abuse before it reaches them — then proves it did, with a tamper-proof receipt on every invocation.
Each capability is available the moment you add Quantum OSCL to your watsonx Orchestrate workspace — just ask in plain language.
Quantum OSCL is designed to support the regulations and standards relevant to security. Confirm applicability with your own compliance team.