Skip to content
Automated Vulnerability Remediation

How the register works, and what it leaves out

The site lists platforms and does not order them by merit. No number rates a platform anywhere in it, and no page names a pick. Vendors appear in alphabetical order on every page that names more than one. Every platform has the same fields in the same order: what it proves about exploitability, how it ships a fix, how it finds an owner, how fast it answers a new disclosure, what agents you can shape, what stack it works on, and what proof it publishes.

A fact enters the register only from a page the editor opened. The entry records the page URL, the date it was read, and a short quotation. A statement on the vendor's own site is labelled the vendor's claim. A statement from press or another outside party is labelled third-party source. Where a page was read and the thing was not there, the entry says none found on that date, with the date. That is a statement about the pages read, not about the product.

This site lists and does not rank

Automated Vulnerability Remediation lists and does not rank. The register gives each of the 14 platforms the same five buyer columns and the same seven topics, sorted by name, and no column adds up to a total or a pick.

How facts are sourced and labelled

A platform's line in the register holds what it claims, who said it, where it was read and when. Vendor's claim marks a platform describing itself, for example a pull request rate on its own blog. Third-party source marks an analyst, a news outlet or a customer case not hosted by the vendor. None found on a date marks a page searched on that day where the point was absent.

  • The read date sits beside every source link on the profile and again in the Sources list under it.
  • First edition pages were read by 10 Nov 2026; the news added on 19 Nov 2026 carries its own read date.
  • A wrong line is fixed on the profile and logged with its date on the corrections page.

Facts read per platform

The counts below are the fact lines recorded for each platform, split by label. They show how much was published and read, not how good a platform is.

Scroll the table sideways for more columns

Platforms in alphabetical order. Counts are the fact lines read for each platform up to 10 Nov 2026; the labels are explained on the Method page.
PlatformWhat it isFacts readThird-party factsVendor claimsNone-found lines
Astelia SecurityAI-native exposure management with reachability analysis and agentic remediation411355
AverlonAgentic remediation (exposure/attack-chain analysis with fixes delivered into developer workflows)400364
Backline SecurityAgentic remediation (Autonomous Exposure Remediation)443383
Cogent SecurityAgentic vulnerability management and remediation (AI agents across the full vulnerability lifecycle)4010282
Dux SecurityAgentic exposure management (exploitability analysis, mitigation, remediation acceleration)280208
MazeAgentic vulnerability investigation and remediation (code and cloud)370334
Onit SecurityAgentic exposure management / remediation (decision-based)371324
QualysExposure/RBVM platform with remediation and agentic AI370361
Rapid7Exposure/RBVM platform with remediation workflows310283
Rival SecurityAgentic security platform / agent harness for vulnerability management, code and cloud security400355
SeemplicityAgentic exposure management and response (RBVM/CTEM workflow with AI analysts)410383
TenableExposure management platform with remediation and agentic orchestration391353
VicariusRemediation-first exposure platform (patching, scripting, patchless protection)380362
ZafranExposure management platform with mitigation via existing controls (agentic features)380362

Order rules

  • Vendors are listed alphabetically by display name, ignoring case. No vendor is placed first by design, and no page orders vendors by merit.
  • Every vendor has the same fields in the same order: what it is, what it does well, where it is limited, then the facts under the seven topics, then profile facts.
  • The two vendors of a side-by-side page are ordered alphabetically too, in the address and in the title, and the list of pairs is alphabetical.
  • News items are listed by date, newest first.
  • Glossary terms are listed alphabetically.

The seven topics

Facts are grouped under seven topics. They are headings for reading, not measures, and no topic carries a weight.

  • Proves exploitability in your environment: whether the product shows that a finding is reachable and exploitable in the customer's own cloud, code and controls, with the evidence.
  • Ships the fix: whether the product produces the fix itself (a tested pull request, a configuration change, a mitigation) with a human approving, or stops at a ticket.
  • Finds the owner: whether it resolves each finding to the team that can fix it, and follows up.
  • Speed on a new disclosure: how fast it answers where a new CVE hits the customer after a public disclosure, as documented.
  • Agents you can shape: ready-made agents, agents the vendor builds for the customer and agents the customer builds, with the controls on each.
  • Works on the stack you have: which scanners, cloud, code hosts and ticketing systems it reads.
  • Proof: published results with a method, named customers and research.

Facts are dated. A fact read on 10 Nov 2026 describes the page on that day. Vendors change their sites, so every entry shows when it was read, and the register is reissued as pages are read again.

Corrections are made on the page and listed with their date on the corrections page. A vendor or a reader who finds a fact that no longer matches its source can write to editor@automatedvulnerabilityremediation.com with the URL and the line.

Customer results, benchmark figures and speed claims are reproduced as the vendor states them, with their method where the vendor gives one. A figure published without a method is marked as such. The register does not add its own measurement to a vendor's number.

The seven topics are fixed: proves exploitability in your environment, ships the fix, finds the owner, speed on a new disclosure, agents you can shape, works on the stack you have, proof. The two guides on reading a claim and on the trial questions show how to test each topic yourself.

Read dates and refresh

Platform pages change faster than the register: pricing pages appear, agents are renamed. Each profile is re-read at least every 90 days, and the register edition on the datasets page moves with it.

Sources

6 sources, each with its read date
  1. Astelia Security: www.astelia.io/platform, read 10 Nov 2026
  2. Averlon: www.averlon.ai/platform, read 10 Nov 2026
  3. Backline Security: backline.ai/product/, read 10 Nov 2026
  4. Cogent Security: www.cogent.com/platform/assess, read 10 Nov 2026
  5. Dux Security: dux.io/, read 10 Nov 2026
  6. Maze: mazehq.com/platform/cloud, read 10 Nov 2026

Last reviewed 10 Nov 2026.