As AI agents reshape how applications run, traditional security models are struggling to keep pace with workflows that extend far beyond the network perimeter. Arcjet has introduced a new capability aimed at addressing this gap. Arcjet introduces Guards to bring application security directly inside AI agent workflows, enabling developers to enforce protections where modern threats actually occur. The launch reflects a broader shift in cybersecurity toward securing execution layers rather than relying solely on network level defenses.
Unlike conventional tools such as web application firewalls or HTTP middleware, which operate at the request boundary, Guards are designed for environments where much of the logic never passes through a traditional request. In agentic systems, workflows often involve tool calls, background jobs, and queue based processing, meaning untrusted inputs can enter at multiple points that remain invisible to perimeter based security controls.
Arcjet introduces Guards to solve this challenge by embedding enforcement directly into application code. Developers can define and apply security rules at the exact point where untrusted input is introduced, whether through tool outputs, external data sources, or workflow steps. This approach allows security policies to operate within the real context of the application, including user identity, session data, and business logic, rather than relying on external inspection layers.
“Security has to live where the code lives. For agentic systems, that means inside the tool calls and workflow steps where untrusted input actually arrives, not at a perimeter that no longer exists,” said David Mytton, CEO at Arcjet. “Guards give developers a way to enforce policy inside the code paths agents use every day the same place the threat model now lives.”
The new feature integrates with Arcjet’s broader application layer security platform, allowing developers to manage protections within the same codebase as their application logic. This eliminates the need for separate security systems and ensures that protections evolve alongside the application itself through standard development workflows such as code reviews.
With Guards, developers can detect prompt injection attempts before they influence model behavior, block sensitive data such as personally identifiable information from reaching external models, and enforce usage limits within AI driven loops. The capability also supports validation of inputs across non HTTP environments, including background processes and asynchronous workflows.
Arcjet introduces Guards as part of a layered security model that complements existing protections such as endpoint shielding and bot detection. By extending visibility and control into execution paths, the platform aims to provide comprehensive coverage across both external interfaces and internal application logic.
The solution is available through Arcjet’s JavaScript and Python software development kits, allowing existing customers to adopt it immediately while offering new developers access through a trial. As organizations increasingly build applications powered by long running AI agents, embedding security directly into workflows is expected to become a critical requirement.
Arcjet introduces Guards to address this emerging need, signaling a shift in how application security is designed for the AI era. By moving protection closer to where decisions and data interactions occur, the company is helping redefine how developers secure complex, autonomous systems at scale.
Source-businesswire
Recommended Cyber Technology News :
- ABS Acquires RMC Global to Boost Cyber Risk Capabilities
- Team Cymru Launches Pure Signal MCP Server for AI-Driven Threat Intelligence
- CSC Pushes .BRAND Domains for AI Era Security Strategy
To participate in our interviews, please write to our CyberTech Media Room at info@intentamplify.com
