Harnesses

Detectify Launches Agentic Tooling for AI Agents to Test Application Vulnerabilities

November 18, 2026 · 2 min read

a computer screen with a bunch of code on it
Chris Ried / Unsplash

Detectify has released Agentic Tooling, a set of tools that allow corporate AI agents to verify if suspected vulnerabilities can actually be used to attack running applications.

The company also introduced the Security Ledger, which is designed to record evidence for each test. These releases target DevSecOps teams who are uncertain about the reliability of AI agent reports. Detectify customers have reported that approximately 80 percent of findings from AI powered static application security testing tools are irrelevant. The company warned that agents may create convincing reports for tests that never actually occurred when under pressure to find answers.

Data from the Detectify Cyber Hygiene Index for the second half of 2026 indicates that 9 out of 10 unresolved critical and high severity vulnerabilities remained unprotected for more than 90 days. Additionally, a study by researchers at Fudan University and Tencent Holdings found that 76 percent of checked static analysis alerts were false positives, requiring developers to spend 10 to 20 minutes investigating each one.

Agentic Tooling provides agents access to testing engines developed by Detectify over a decade, including community white hat hacker payload evaluations, attack surface data from Surface Monitoring products, and Alfred, an AI system launched in March 2025 for automated known vulnerability testing. These tools operate within the customer environment and connect to agents via the Model Context Protocol or a command line. They process only customer provided data and are said to work with any agent the customer already uses.

The design centers on Loop Engineering, an emerging practice of running AI agents in iterative automated workflows instead of manual prompting. Loops are triggered by events such as continuous code integration, code scan results, bug bounty reports, red teaming, or Detectify attack surface data. Each finding is converted into a ticket. When an agent picks up a ticket, it reads the relevant code before testing the actual application from the outside like an attacker. This continuous loop is described by Detectify as a verification factory.

All requests, payloads, and results are recorded in the Security Ledger whenever an agent uses Detectify tools. Because the ledger is written by Detectify rather than the agent, teams can distinguish findings supported by actual tests from unsupported AI conclusions. Tickets that could cause harm, such as denial of service tests, are marked as unsafe for human decision making.

Agentic Tooling is available upon request and builds upon the Detectify MCP Server announced in May. The company is backed by investors including Insight Partners, which became a majority shareholder in October 2024, along with Balderton Capital, Inventure, First Fellow Partners, and SquareOne App.