iFinder AI agents identify 84 vulnerabilities in 5G core implementations

The discovery of 84 flaws via automated agents signals a shift where vulnerability identification now moves at machine speed, far outpacing manual remediation cycles. For streaming infrastructure relying on 5G, this highlights a growing risk in cloud-native cores where internal interfaces are increasingly exposed through configuration errors or protocol tunneling. As these automated 'agent swarms' become more accessible, the industry must move toward automated defense loops to protect subscriber data sessions from hijacking. Watch for whether major telecom vendors integrate similar multi-agent testing into their CI/CD pipelines to close the gap between discovery and patching.

Nanyang Technological University's iFinder system enters a competitive field where telecom security testing is increasingly automated. In March 2025, the 3GPP Security Assurance Group published updated specifications for 5G core network security testing, establishing baseline requirements that open-source implementations like Open5GS and free5GC must meet before commercial deployment. Meanwhile, the OpenSSF launched its Alpha-Omega project targeting critical open-source infrastructure security, which directly funds security audits of projects that form the backbone of telecom and cloud-native stacks, including several 5G core implementations that AI agents uncover 84 5G core security vulnerabilities tested.

For related background, see StreamingMeme's prior coverage of Multiverse Computing launches Quasar 438B reasoning model for enterprise agents.