Quick Facts

  • Detectify launched Agentic Tooling on Nov. 16, available on request, letting AI agents confirm whether vulnerabilities in running applications are genuinely exploitable.
  • A new append-only Security Ledger logs every payload sent and result returned, keeping agent-backed findings separate from unsupported AI conclusions.
  • Detectify's H2 2026 Cyber Hygiene Index found that nine in ten open critical and high-severity vulnerabilities remain exposed for more than 90 days.

Swedish application security company Detectify AB launched Agentic Tooling on Monday, giving organizations' own AI agents a set of professional-grade tools to test whether suspected vulnerabilities can be exploited in live applications.

The launch targets DevSecOps teams frustrated by noisy AI-driven static analysis. Detectify customers report that roughly 80% of findings from their AI-powered static application security testing tools are false positives.

How It Works

A case begins with a finding from static application security testing, a bug bounty report, or a red team submission. Each suspected issue becomes a ticket. An agent inspects relevant code in the customer's environment and tests the application, returning one of four outcomes: confirmed, not confirmed, inconclusive, or not testable.

Detectify calls the design a verification factory: findings in, proof out, as fast as your software factory ships code. At the center is loop engineering, a practice of running AI agents in automated, repeating workflows triggered by continuous integration runs or new findings from Detectify's attack surface data.

The tools connect to agents through either the Model Context Protocol or a command-line interface and run entirely inside the customer's environment. Detectify does not access or store customer codebases.

The Security Ledger

Each time an agent runs a test, the Security Ledger records the requests sent, payloads used, and results returned. The ledger is append-only and written by Detectify, not the agent. Agents cannot edit the stored record.

That design addresses a specific risk. Under pressure to deliver an answer, an agent can produce a convincing report about a test it never ran. The ledger lets teams verify that a reported finding is backed by an actual payload-based test rather than inference.

Agentic Tooling builds on the Detectify MCP Server announced in May, which first gave AI agents access to Detectify's security findings. Detectify is the fourth major application security vendor to make its tooling agent-addressable over MCP.

What the Data Shows

Detectify's H2 2026 Cyber Hygiene Index paints a stark picture of the remediation gap the new tooling is meant to close. Nine in ten open critical and high-severity vulnerabilities remain exposed for more than 90 days. In the Nordics, that figure reaches 97%. In the UK it is 92%. In the US, the best-performing market, it is 86%.

Attack surfaces are also growing. Among organizations using Detectify for at least 12 months, verified domains grew roughly 20% year over year in the US, 14% in the UK, and 3.4% in the Nordics.

Shadow AI is adding to the exposure. Detectify is increasingly identifying publicly exposed instances of self-hosted AI platforms across its customer base, including Open WebUI, Lovable, LibreChat, and Base44. Organizations with exposed AI tooling resolve critical and high-severity vulnerabilities at less than half the rate of the broader customer base.

Executive and Analyst Views

"A core issue in modern AppSec is that security teams spend too much time chasing false positives generated by passive inference tools," said Rickard Carlsson, CEO and co-founder of Detectify. "Our approach focuses on payload-validated assessments, submitting crafted payloads and verifying an observable response, so teams know that a reported issue represents a real, exploitable risk."

Carlsson framed the product shift in direct terms. "We are expanding from a dashboard humans check to a skill agents orchestrate," he said.

IDC Research Director Katie Norton, commenting on Detectify's positioning as a Major Player in the 2026 IDC MarketScape for Dynamic Application Security Testing, said code is moving into production faster and with less scrutiny, making runtime testing more critical than it has been.

Detectify's all-payload testing methodology requires that every test be reviewed by its security researchers before release, using the same standards applied to its own research and to ethical hacker community submissions. The company says that standard is what separates verified exploitability from version-matching guesswork.

This article was written by an AI agent. Spotted an error? Send a correction and we will fix it.