How it works

ReconForge runs security assessments on its own. You choose a ready-made template, point it at a target, and launch a run. From there an AI agent drives real security tools, gathers evidence as it goes, and hands back findings ranked by real-world risk, each with the proof behind it and clear steps to fix it.

The shape of an assessment

Three ideas carry the whole product:

• You pick a template, you do not script a test. A template is a ready-made assessment type, from a quick surface check to a deep review. You choose one instead of configuring tools by hand.
• An agent does the work, with evidence. The agent decides what to run, runs it, and records every command and its output. Nothing in a finding is a guess. It traces back to something that actually ran.
• Results are ranked by risk, not raw output. You get findings ordered by real-world impact, with remediation, ready to share with engineering or hand to an auditor.

Two phases: discovery, then confirmation

An assessment has up to two phases:

• Discovery finds your exposure and surfaces weaknesses. It produces candidate findings, which are detected but not yet proven.
• Confirmation (exploitation) safely proves which candidates are genuinely exploitable, and upgrades them to confirmed with proof-of-concept evidence. This phase is available on self-hosted deployments.

The managed cloud runs full discovery against domains you have verified you own. Self-hosted runs everything, including confirmation and internal-network testing, entirely inside your own environment. See How an assessment runs for the detail.

Why this is different

A raw scanner tells you something might be wrong. A manual test proves it but takes an expert and time. ReconForge aims for both at once: the breadth of automation with the evidence of a real test, produced without you scripting each step.

Next

• Core concepts: the words you will see everywhere.
• How an assessment runs: discovery, confirmation, and the live view.
• Roles and access: who can see and do what.