Breach and attack simulation
Breach and attack simulation runs safe, scripted attack scenarios against your own controls and records which ones were blocked, which were detected and which got through unseen.
What it proves
A BAS run produces a control-by-control record. The email gateway let three of twenty malicious attachments through. The EDR blocked credential dumping on Windows hosts but only alerted on the macOS fleet. The SIEM received the events but no rule fired.
Because the scenarios are scripted and repeatable, the same test can run every week. That turns a one-time finding into a trend line, and it shows when a policy change or an agent update quietly breaks a control.
What it does not prove
BAS does not tell you whether an attacker can get from a phished laptop to your domain controllers. Each scenario tests a technique in isolation, usually from an agent you placed. It assumes the attacker is already where the agent sits.
It also says little about vulnerabilities you have not patched. A BAS tool tests behaviour against controls, not the exploitability of a specific CVE on a specific server.
How it differs from its neighbours
Adversarial exposure validation is the wider category Gartner introduced in 2025, and it absorbs BAS. Most BAS vendors now sell under the AEV label. The practical difference: an AEV platform adds attack-path testing or exploit validation next to the control tests.
Automated pentesting starts from the other end. It looks for a way in and through, and treats the controls as obstacles. BAS treats the controls as the subject.
Who buys it
SOC managers and detection engineers, usually after a tool consolidation or an EDR migration, when someone asks whether the new stack actually catches what the old one did. Compliance teams under DORA or NIS2 also buy it for repeatable evidence of control testing.
Expect to install agents. Most BAS products on this site need a test point, simulator, validator or actor on the hosts they test from. AttackIQ Flex is the documented exception: agentless test packages, Windows only, with a smaller content set.
Vendors that fit
Listed alphabetically: documented first, then partial claims. The note under each name is a summary of what the vendor's own pages claim, as read on 3 November 2026.
Documented (7)
- AttackIQ (documented)AttackIQ's AWS listing and Flex FAQ describe breach and attack simulation; DISA's selection names it an AEV platform; the Exposure Management Module adds attack-path modeling and risk scoring; Enterprise is described as a 'prescriptive service' run with AttackIQ experts (managed/co-managed); 'offensive testing' of full attack paths is marketed on the homepage, but no stand-alone automated pentest product was found.
- Cymulate (documented)Cymulate's AWS listing describes an exposure validation platform 'that includes breach and attack simulation and automated red teaming'; the product whitepaper documents Attack Path Discovery (lateral movement via Pass-the-Hash, Kerberoasting, password spraying) and CTEM prioritization with a validated exposure score; no dedicated 'automated penetration testing' product or managed validation service was found on the pages read.
- Mandiant Security Validation (documented)Google Cloud describes the product as testing security controls with breach and attack emulations; continuous drift testing (AEDA) supports part of the exposure-validation use case, but no attack-path or automated pentest capability is claimed on the product page.
- Picus Security (documented)Picus's July 2026 launch release names three products (Breach and Attack Simulation, Autonomous Penetration Testing, Exposure Validation) under one AEV platform; the APV datasheet documents attack path mapping to domain admin; EXV ingests scanner findings (Tenable, Qualys, Wiz) and Attack Surface Validation adds CAASM; Picus sells 100% through channel partners including MSSPs, but no Picus-run managed service was found.
- SafeBreach (documented)SafeBreach's own pages call Validate a BAS tool and Propagate 'an enterprise-grade automated penetration testing tool' for attack path validation, and brand the combination as AEV; the CTEM Platform adds Helm agents for scoping, discovery and prioritization but relies on third-party VM and EASM data, with its own Exposure Hub announced as 'upcoming' in April 2026; no SafeBreach-run managed service was found (a Ness professional-services offer on AWS Marketplace is third party).
- SCYTHE (documented)SCYTHE's own pages position it as Adversarial Exposure Validation that evolved from BAS (5.0 release) and list Managed AEV and Managed Purple Teaming services on the pricing page; the AI campaign builder generates attack sequences, but no attack-path graph or exposure-management module is claimed.
- Validato (documented)Validato describes itself as a BAS / continuous security control validation platform and its FAQ describes it as using Adversarial Exposure Validation; it says it uses simulation of ATT&CK behaviours rather than payload emulation, and no attack-path, pentest or managed service is claimed.
Partial (1)
- Pentera (partial)Pentera's own pages describe AI-driven internal, external and cloud pentesting (Core, Surface, Cloud), attack path root-cause analysis and Resolve remediation workflows supporting CTEM, and it was named a 2026 Gartner Peer Insights Customers' Choice for AEV; security control assessment (EDR, SIEM, firewalls) happens inside attack runs rather than through a separate BAS scenario library; human-led red-team services come via the 2025 EVA acquisition rather than a managed validation subscription.