Unbound AI has officially introduced the Agent Access Security Broker (AASB), establishing a new cybersecurity category designed to address the growing risks associated with AI coding agents. Alongside this category launch, the company unveiled its AASB platform, which enables enterprises to discover, monitor, and govern AI coding agents across development environments. As AI-powered tools rapidly become central to software development workflows, organizations are now under pressure to balance innovation with security, compliance, and operational control.
Today, AI coding agents such as Cursor, Claude Code, Copilot, and Codex are transforming how developers build and deploy software. These tools can write code, execute terminal commands, provision infrastructure, and interact with APIs and external systems. However, while they significantly improve productivity, they also introduce a new and complex attack surface. Traditional security solutions including AppSec, IAM, CASB, and endpoint tools were not designed to manage autonomous agents operating within live development environments. As a result, enterprises often lack visibility into how these agents behave, what permissions they hold, and how they interact with critical systems.
Consequently, security teams face a growing governance challenge. Many organizations cannot clearly identify which AI agents are active, how they are configured, or what actions they are performing. Without proper oversight, these agents may operate with high-level permissions and minimal supervision, increasing the risk of misconfigurations, unauthorized actions, and compliance violations. Therefore, Unbound AI developed AASB as a dedicated control layer to close this gap.
AASB functions similarly to how Cloud Access Security Brokers (CASBs) transformed cloud security. However, instead of governing human access to SaaS platforms, AASB governs AI-driven interactions across development ecosystems. It creates a control and enforcement layer between AI coding agents and the systems they access, including IDEs, terminals, infrastructure, APIs, databases, and external tools.
“CASB was built for a world of human access to SaaS,” said Raj Srinivasan, CEO of Unbound AI. “AI coding agents changed the problem. Enterprises now need to govern software that can read, write, execute, connect, and act with enterprise permissions. We created the AASB category because the industry needs a control plane for agent access before the first destructive command, unsafe MCP action, or compliance gap forces the issue.”
Through its platform, Unbound AASB enables organizations to discover AI agents, track configurations, and identify risky behaviors such as excessive permissions or auto-approval settings. Additionally, it allows teams to audit, block, or require approval for high-risk actions, including destructive terminal commands and unsafe external integrations. At the same time, the platform generates audit-ready reports to support compliance and internal governance requirements. Importantly, it introduces these controls without disrupting developer workflows, allowing organizations to maintain productivity while improving security oversight.
“Security leaders do not need another reason to say no to AI coding agents,” Srinivasan said. “They need a way to say yes safely. Unbound lets organizations keep the productivity gains of AI coding tools while giving security and compliance teams visibility, policy, approvals, and evidence over the highest-risk actions.”
The urgency for such a solution is increasing rapidly. Industry data shows that AI adoption in development environments is accelerating, with a large percentage of developers already relying on AI tools. At the same time, many of these tools operate outside formal governance frameworks. Moreover, studies have revealed widespread use of insecure configurations, including long-lived credentials and publicly exposed integration servers. As AI agents become more autonomous, the risks associated with uncontrolled actions also grow significantly.
Real-world incidents further highlight the need for governance. In one case, an AI coding agent tasked with fixing a minor issue reportedly triggered a major infrastructure disruption by executing unintended actions. While such events may stem from configuration issues, they underscore the lack of policy enforcement layers capable of preventing high-impact mistakes.
Unbound AI positions AASB as the next evolution of security control planes, complementing rather than replacing CASB. While CASB continues to govern human interactions with cloud services, AASB focuses specifically on agentic software behavior, including runtime actions, configuration states, and system-level access. This distinction is critical as enterprises move toward increasingly autonomous development environments.
In early deployments, Unbound’s platform has already identified and prevented multiple instances of excessive or unintended agent activity, including unauthorized code changes, service restarts, and connections to unsanctioned tools. By providing visibility, enforcement, and human-in-the-loop controls, the platform enables organizations to manage AI-driven development safely and effectively.
Looking ahead, Unbound AI believes AASB will become a foundational component of modern software security architectures. As enterprises continue to scale AI adoption, the need for dedicated governance layers will only intensify. Through this launch, Unbound is positioning itself at the forefront of securing the next generation of AI-powered software development.
Recommended Cyber Technology News:
- Sensitive Data of 2.7M Users Exposed in Navia Breach
- BigID Achieves FedRAMP Certification for AI Data Security
- Saviynt and Cyera Partner to Unite Identity and Data Security
To participate in our interviews, please write to our CyberTech Media Room at info@intentamplify.com
