Breaking News

Solo open-source projects address challenges of agentic AI

https://ift.tt/tIBcMrJ

The rush to adopt agentic AI presents significant challenges for enterprises, particularly around governance, security, and ensuring reliability in production environments. Two new open-source projects created by Solo.io, Agent Registry and Agent Evals, aim to solve these critical adoption hurdles.

The Agent Registry was open-sourced at KubeCon Atlanta and subsequently donated to the Cloud Native Computing Foundation as a sandbox project in AmsterdamIts creation addresses the enterprise need to curate and govern approved AI agents, MCP tools, and Agent skillsBy serving as a central hub for hosting these artifacts, it provides governance, intelligent searching, and allows developers to easily build, push, and run agents in environments like KubernetesThe agents themselves are highly customizable, supporting multiple frameworks including the declarative YAML-based Key Agent, as well as Agent Core, Azure, and Google ADKCustomization allows users to configure agent instructions, skills, MCP tools, and model settings. That, said Lin Sun, the director of open source at Solo.io, which created the projects, is how and why Agent Registry was built.

“As we were running agents at Solo, we use Kagent a lot to help us troubleshoot Kubernetes environment deployment issues, networking configuration issues. Because they are not deterministic, some agents are a little bit more reliable with certain models with certain prompts,” Lin said. “So we feel there’s a strong need to be able to ship agents with reliability and confidence in mind.” 

A separate project, Agent Evals, was announced to enable the reliable shipping of agentsThis project was born out of internal experience where agents were found to be non-deterministic, creating a strong need for reliability and confidenceAgent Evals provides tooling to benchmark agents by leveraging open standards like OpenTelemetryIt collects real-time metrics and tracing as the agent runs to score performance and inference quality, producing a report that helps users understand their agent’s reliabilityThis assessment is crucial for determining the level of human intervention required, whether fully autonomous, human-in-the-loop, or human-outer-loopAgent Evals works in conjunction with other observability tools that support OpenTelemetry standards.

Moving beyond individual developer laptops into full production requires robust security and governanceSolo is addressing this by solving problems such as securing agent communication with LLMs and MCP toolsThe Agent Gateway provides a critical solution, offering centralized policy, enforcement, security, and observability for trafficThis includes “context layer enforcement,” which can be configured to put guardrails on responses, for instance, stripping out sensitive data like credit card or bank account numbers as traffic travels through the gatewayFurthermore, Agent Gateway is being integrated into Istio as an experimental data plane option in Istio Ambient mode, helping mediate agent traffic without requiring changes to the agents or MCP tools themselves.

Collectively, these tools—Agent Registry for governance, Agent Evals for reliability, and Agent Gateway for security—are filling in the puzzles needed to run agentic AI in production with confidenceHowever, for critical work, human involvement remains a necessary component, as the philosophy suggests viewing the agent like a growing co-worker that still benefits from supervision and peer review.

“I’m always thinking about the agent as like a person,” Lin told SD Times. “Even with your coworker, you don’t always trust their work. You need a peer review of the work, to iterate and make it better. So, at this stage of the agent, maybe it’s more like from toddler to kindergarten. It’s growing, right? But even when the agent becomes an adult, like my son just turned 18, you still need to kind of supervise a little bit of providing some insights.”

 

The post Solo open-source projects address challenges of agentic AI appeared first on SD Times.



Tech Developers

No comments