search
open protocol
Trends
- 1MCP protocol flagged as major security risk for AI agents●MCP for agent-to-agent comms may be the riskiest protocol you've never heard of
A vulnerability in AI agents from Google and other companies has exposed what researchers describe as a structural flaw in the Model Context Protocol (MCP), a widely adopted standard for connecting AI agents to tools and data. Security coverage argues MCP was never designed for agent-to-agent communication and that its architecture leaves systems open to prompt injection and privilege abuse, prompting debate about whether the industry adopted the protocol too quickly without adequate safeguards.
- 2Open-Source AI Agents Win the Enterprise Standards War●The Open-Source Enterprise Agent Offensive — How Open Models, Protocols, and Governance Are Winning the Standards War
A new analysis argues that open-source AI agents are gaining ground in enterprise settings, with open models, open protocols, and community-driven governance giving them an edge over proprietary alternatives in the emerging standards battle. The piece frames the shift as an 'offensive' by the open-source ecosystem, suggesting companies increasingly favour transparent, interoperable agent stacks over closed platforms.
- 3Open-source EEG toolbox aims to make emotion recognition more reliable●Let machines read emotions more reliably: open-source EEG toolbox helps benchmarking evaluation under standardized protocols
Researchers have released an open-source EEG toolbox designed to help machines read human emotions more reliably. The software enables benchmarking of emotion-recognition algorithms under standardized protocols, addressing inconsistent evaluation practices in the field. The release, announced via EurekAlert!, is expected to make it easier for labs to compare results and reproduce studies in brain-signal-based emotion detection.