A startup called Vectara Inc. wants to help enterprises safeguard themselves against any potential damage or problems caused by artificial intelligence agents, and its solution is yet another agent.
The company today announced the immediate availability of its new Tool Validator agent, the second in its family of “Guardian Agents.” They’re designed to enforce governance and security policies across all agentic systems in real time.
It has become well apparent to many companies that agentic AI systems are struggling to live up to their potential and unable to deliver the promised Golden Age of enterprise automation due to their brittle nature, the tendency for workflows to break down and some alarming errors that have cropped up in recent weeks.
Vectara says many agentic AI breakdowns result from errors in the action plans that AI agents create before they go about completing a task. If an error is made at the very beginning of an agent’s workflow, it will compound as each following action in the chain is processed, resulting in low-quality or totally irrelevant and confusing outputs, and sometimes extremely disastrous consequences. “While 2025 began with tremendous enthusiasm for the promise of AI agents, many organizations have since come to realize it is almost impossible for an agentic system to deliver on its promises or deliver the ROI that enterprises demand,” said Vectara co-founder and Chief Executive Amr Awadallah.
Preventing agentic annihilation
A case in point involves Google LLC’s newly launched Antigravity platform, which is an integrated developer environment that embeds agentic functionality at its core in an effort to automate as much of the software development workflow as possible. Earlier this month, a photographer and graphic designer from Greece was experimenting with Antigravity to create a new image selection application, which would enable him to rate individual photos and automatically sort them into different folders based on that rating.
While creating the app with the assistance of Antigravity’s AI agents, the platform advised him of the need for a server restart, which also required the project’s cache to be cleared. The developer assumed that Antigravity’s AI agent would be more than capable of doing this, so he instructed it to complete the task – but instead of clearing the cache, it initiated a command that wiped his entire disk drive from existence.
An incident overview posted on GitHub makes it clear that the unfortunate accident was the result of an incorrect tool being used by the agent responsible. That’s exactly the kind of situation that Vectara’s new Tool Validator is supposed to prevent.
Vectara said Tool Validator would have stopped Antigravity’s agent from wiping the developer’s hard drive because of the way it eliminates all unnecessary calls by third-party agents. It acts like an agentic overseer, preventing agents from accessing tools such as web browsers, databases and file systems when they don’t actually need to use them. It leverages Vectara’s existing Agent Operating System, which combines multimodal retrieval-augmented generation with orchestration and always-on governance to create trust in AI systems and their actions.
When an agent is assigned a task, it will generate a list of planning steps, and this is where Tool Validator comes in. Its job is to review the list and check that the agent’s proposed tools are actually necessary for it to complete the assigned task. If it sees any irrelevant tools being suggested, it will flag this to the agent and prevent it from being used. Tool Validator will also recommend missing or alternative tools to the agent, so it can better execute each workflow with full traceability, the company said.
Tool Validator is designed to work in concert with Vectara’s existing Hallucination Corrector Guardian Agent. It reviews an AI agent’s outputs and tries to identify hallucinations as they attempt to use resources and tools outside of the main RAG pipeline.
Image: Vectara
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.
About News Media
Founded by tech visionaries John Furrier and Dave Vellante, News 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.
