Dublin – October 07, 2026 -- Edgescan has launched Edgescan Atomic, an autonomous penetration-testing capability built on agentic AI, designed to close the gap left by scheduled security assessments. The tool can run on top of Edgescan's existing continuous security platform or operate as a standalone agentic penetration-testing solution.
Atomic replaces fixed testing schedules with on-demand validation
Traditional penetration-testing programs rely on periodic, scheduled engagements, leaving organizations exposed between tests as application code changes and threats evolve. Edgescan Atomic allows security teams to launch autonomous testing immediately, using previous target metadata when available, with validation handled by an agentic team backed by human experts when required.
Platform draws on historical scan and vulnerability data to sharpen attack-path discovery
Atomic becomes more effective when integrated with Edgescan Intelligence, which includes historical scan data, PTaaS results, false-positive records, validated vulnerabilities, asset context, authentication workflows, API information and prior assessment history. This context gives the AI a richer starting point to explore and validate potential attack paths rather than testing blind.
Edgescan Harness imposes scope, permissions and audit controls on AI actions
The company pairs its security intelligence with an enforcement layer called the Edgescan Harness, which controls what the AI can and cannot do. The Harness enforces defined scope, permissions, approvals and policies, letting Atomic reason and adapt like an attacker while keeping its actions controlled and auditable -- a structure intended to let organizations adopt autonomous testing without losing oversight.
Tool targets business-logic flaws while cutting analyst triage workload
Edgescan positions Atomic to identify business-logic and access-control vulnerabilities, including insecure direct object reference (IDOR), authorization weaknesses and broken access control. Findings are validated with evidence before reaching analysts, a design meant to reduce investigation and triage time for security operations and AppSec teams. The licensing model is structured to provide a defined testing budget, with the Harness logging and policy-enforcing all AI actions for cost predictability.
CEO cites AI-driven threats and code-change velocity as the catalyst
Eoin Keary, CEO of Edgescan, said Atomic addresses a fundamental limitation of traditional penetration testing: the AI-powered threat environment and the pace of application code change. He said the product puts control, flexibility and faster security validation directly in customers' hands.