MOUNTAIN THEORY VS WIDEFIELD SECURITY (CISCO)
Cisco moved to acquire WideField in June 2026 to strengthen Splunk's Agentic SOC, adding identity, credential and session intelligence to threat investigation. It is the third leg of the same strategy that brought in Astrix, and it sits upstream of where Mountain Theory operates.
| WideField Security (Cisco) | Mountain Theory | |
|---|---|---|
| What it controls | Identity lifecycle, credentials, sessions and blast radius | The action an AI agent is about to take |
| Where it sits | In the SOC, during investigation | Inline at execution, between the decision and the action |
| How policy is set | Detection and analytics rules | Plain English, no code |
| Deployment reach | Human, machine and AI identities, strongest inside Splunk | Model and framework agnostic, including custom and on-prem agents |
| Best fit when | Your SOC needs to understand agentic incidents faster | An AI acting wrongly has physical or regulatory consequences |
Why you might pick WideField Security (Cisco)
WideField gives continuous visibility across human, machine and AI identities, with the analytics to spot both external and insider threats. Inside Splunk it makes an investigation far richer: who the identity was, what credentials it held, what session it ran in and how far the blast radius reached. If you already run Splunk, this arrives as part of the platform rather than as a new vendor.
Why you might pick Mountain Theory
WideField expands what an investigation can see. Mountain Theory decides whether there is an incident to investigate. Identity and session intelligence tells you the shape of what happened, in detail, once it has happened. Mountain Theory checks the action against policy before it executes, so the same event becomes a blocked line in a log rather than a blast radius to map.
The honest verdict
These are sequential, not competing. WideField makes your SOC better at understanding an agentic incident. Mountain Theory reduces how many of them reach the SOC. If your investigation queue is the pain, WideField is the answer. If the pain is that the action already ran, no amount of investigative depth changes that.
What we can actually show
Claims in this category are easy to make and hard to check, so here is ours on the record. The same 10 actions were run in the same order under three configurations. Ungoverned, 10 of 10 executed. Under NVIDIA OpenShell alone, all 5 sandbox-boundary crossings were denied at the kernel, and all 3 in-bounds bad decisions still went through, including a secrets read that printed credentials to the screen. Under OpenShell plus Mountain Theory, those same 3 actions returned HOLD, HOLD and BLOCK, and the secrets read was stopped before it executed, so the credentials never printed. Terminal recordings of all three runs are published, including the two actions Mountain Theory has no policy for.
Separately, when a third-party provider updated the foundation model driving an autonomous agent, the agent began attempting multi-step actions it had never tried before. Nothing on our side changed. Every attempt was stopped on 30 and 31 July 2026, the days the behaviour first appeared. No new rule, no signature, no patch.
Watch the three-configuration run against NVIDIA OpenShell
See novel agent behaviour stopped the day it appeared
Ask WideField Security (Cisco), and every other vendor you are evaluating, for the same four things: the exact action set, the ungoverned control condition, the outcome per action including the ones the product did not stop, and the recording. A certification, an integration list or a customer logo answers a different question.
Compare all 56 AI security vendors