Adversarial exposure validation fits your answers
You need both kinds of evidence: that controls hold, and that exposures can or cannot be chained into a real attack, run often and by people who will act on it.
Why this type matched
You asked to prove both control performance and exploitable paths, which is the definition of the category. AEV platforms sell control tests and attack-path or exploit tests under one licence, so one pilot covers both questions.
These answers add points to adversarial exposure validation. If you came from the finder, yours are marked.
- Question 1, “Our controls block and detect attacks”: +1
- Question 1, “Both”: +3
- Question 2, “Hybrid”: +1
- Question 3, “Yes”: +1
- Question 3, “On some hosts”: +1
- Question 4, “An in-house red team”: +2
- Question 5, “Continuously”: +1
What to check in a pilot
Tick these off with the vendor in the room. Print this page; the print layout drops the site chrome.
Vendors that document it
Alphabetical. One factual line each, from the vendor's own material or marketplace listing.
Documented (6)
- AttackIQ (documented)Runs ATT&CK-mapped emulations, from a free agentless Flex tier (Windows only) up to a quarterly Enterprise subscription; Flex prices are published.
- Cymulate (documented)SaaS with one test point per environment, testing endpoint, email, web gateway, WAF, cloud and AD controls; one AWS package is listed at USD 13,000 per 12 months.
- Pentera (documented)Agentless automated pentesting in live production across internal, external and cloud estates; one AWS contract is listed at USD 120,000 per 12 months.
- Picus Security (documented)BAS, autonomous pentesting and exposure validation in one platform, as SaaS, on-premises or air-gapped; AWS entry bundle at USD 150,000 per 12 months.
- SafeBreach (documented)Validate (BAS) plus Propagate (attack path validation) from deployed simulators, including air-gapped and OT sites; no public prices.
- SCYTHE (documented)Agent-based adversary emulation sold with Managed AEV and Managed Purple Teaming services; SaaS, on-premises or air-gapped; quote only.
Partial (4)
- Horizon3.ai NodeZero (partial)Agentless autonomous pentesting run from a Docker host or OVA; AWS packages start at USD 25,000 per 12 months for 500 assets.
- Mandiant Security Validation (partial)A Director plus Actors (agents) run Mandiant frontline content against firewalls, proxies, DLP, EDR and SIEM; no public price.
- Validato (partial)Agent-based BAS of endpoint behaviours aimed at DORA and NIS2 evidence; edition prices from USD 20,000 a year, vendor-supplied on G2.
- XM Cyber (partial)Models attack paths in a digital twin instead of running attacks; AWS lists USD 350,000 per 12 months; CrowdStrike agreed to buy its IP on 16 July 2026.
Close neighbours
If the result surprised you, read the type next door: breach and attack simulation or automated pentesting. The full note on adversarial exposure validation explains what it does not prove.