Harness launches Agent DLC for developers to deploy AI agents using familiar processes and tools

Harness launches Agent DLC for developers to deploy AI agents using familiar processes and tools Integrated software delivery platform provider Harness Inc. said today it’s reinventing the development lifecycle for artificial intelligence agents so organizations can bring them from concept to production without worrying they might break something. The company has announced the availability of Agent DLC, which provides companies with evaluation, deployment governance and security controls that are purpose-built for the uniqueness of AI agents and the strange ways they behave. Crucially, it enables enterprises to develop and manage their agents using the same continuous integration and continuous deployment pipelines they use to ship traditional software code. Harness says companies urgently need to overhaul their agentic development pipelines. At most big companies, senior executives are urging their software teams to adopt AI agents that can autonomously carry out business work on behalf of humans with minimal oversight, anticipating a massive leap in productivity. That’s the promise, but AI agents will only deliver on such gains if they don’t go haywire, and there are very real concerns that they will. Harness cites internal data that shows how only 8% of organizations have gotten agentic AI into production so far, despite all of the hype around the technology. That’s because everyone is struggling to operationalize agents in a secure way. What they’ve found is that the traditional tools used to ship deterministic code simply aren’t up to the job, because AI agents have to make different decisions every time they’re asked to do something. With traditional software, developers can simply test their code once and they’ll be pretty confident that it will deliver the same result every time it runs. But AI agents are an altogether different beast. Each time they’re prompted to do something, they have to decide autonomously which application programming interfaces, tools and sequence they’ll use to accomplish the task at hand. But it means they behave in unpredictable ways, and so the standard software testing tools simply aren’t enough to ensure they won’t cause problems. Integrated security and visibility for AI agents Harness says the answer is to integrate purpose-built guardrails into developer’s existing workflows, rather than try to reinvent the wheel. Agent DLC gives teams the ability to build, test, deploy, operate and govern AI agents through the same platforms they use for their existing applications – with the same, familiar controls, pipelines and governance rules. What Agent DLC does is integrate specialized components throughout the software development lifecycle. They include Harness AI Evals, which makes it possible for developers to define evaluation datasets and create quality gates that run each time an agent or model changes, in order to catch regressions. For deploying new agents and updates, Harness Agent Deployments adds support for managed agent runtimes with third-party continuous delivery platforms such as Amazon Bedrock AgentCore, so developers don’t have to replace their standard release processes. On the operational side, AI Configs support release and prompt management and model changes at runtime, so instant rollbacks can be initiated without having to redeploy previous versions. There’s also a new asset catalog integrated within the Harness Internal Developer Portal that can automatically discover and register agents, skills and plugins to prevent tool sprawl and duplicate work. Harness is taking security and governance seriously too. The Agent Security module within Agent DLC is designed to scan the AI models and skills that agents leverage to try and identify misconfigurations. It can also generate a “bill of materials” for each agent and test for adversarial inputs before deployment, Harness said. In addition, it acts as a firewall for agents in production, continuously forcing security policies to prevent prompt injection attacks and data exfiltration. The final piece of the puzzle is Harness AgentTrace, which is designed to trace AI agent’s actions as they work and enable comprehensive auditing of whatever they get up to, the company said. With AI agents, the possibility exists that two agents might produce the same outcome while taking drastically different paths. AgentTrace captures this execution process at the run level and across full sessions, generating telemetry that can help developers to identify performance bottlenecks and compare the quality of their outputs. The foundational components of Harness AgentTrace are being made available through an open-source license to encourage widespread adoption, Harness said. Harness said none of these capabilities are actually novel solutions that didn’t exist before, but until now, enterprises have always had to implement them separately. That means stitching together and maintaining separate evaluators, runtime firewalls, security scanners and traces, which is a massive drag on time and resources. By unifying all of these capabilities, Agent DLC gives developers a much more elegant and simplistic model for deploying AI agents, with access controls, audit trails, security and governance all rolled into the same development lifecycle. Image: SiliconANGLE/Microsoft Designer A message from John Furrier, co-founder of SiliconANGLE: Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network. Are you AWS customer? Support SiliconANGLE Financially by buying your AWS services from our Marketplace portal page and links. About SiliconANGLE Media SiliconANGLE Media is a recognized leader in digital media innovation, uniting breakthrough technology, strategic insights and real-time audience engagement. As the parent company of SiliconANGLE, theCUBE Network, theCUBE Research, CUBE365, theCUBE AI and theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI. Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.