Shadow MDM
Device management that proposes, explains and waits for your approval
Shadow MDM is lightweight, AI-native device management built from scratch in Rust inside the Shadow agent you already run. There is no extra MDM agent, no osquery and no Fleet. It is suggest-first: small on-device models learn each laptop and write Proposals, an admin approves them in Shadow Hub or Meridian, and the change rolls out to a pilot group before everyone, with one-click rollback. Nothing changes without approval.
Shadow MDM is included in Shadow Hub at no extra cost, along with posture checks.
How it works
Section titled “How it works”Shadow MDM runs one loop, from what a laptop looks like today to evidence your auditor can check.
| Step | What happens |
|---|---|
| 1. Sense | The Shadow agent checks each device: disk encryption, screen lock, firewall, OS updates and patch level, anti-malware, admin rights, password policy, removable storage, and the AI tools, extensions and local model runtimes on it. |
| 2. Understand | Small models on the device learn its idle windows, the performance cost of a change, how sensitive the data it sends to AI is, and whether it is drifting (for example, updates deferred again and again). |
| 3. Propose | Shadow Hub turns gaps into Proposals, ranked by risk, deadline and reach. At most ten open at a time, so the queue stays readable. |
| 4. Approve | An admin approves, narrows the scope, schedules it or declines with a reason (recorded as a risk acceptance). In Shadow Hub or Meridian. |
| 5. Roll out | About 10% of target devices first, least sensitive first, for a pilot window. If any pilot device fails, crashes or a user reverts the setting, the rollout pauses. Then everyone. One click rolls back. |
| 6. Prove | The Enforcement Map links every enforced setting to the controls it satisfies, with who approved it and when. |
What a Proposal looks like
Section titled “What a Proposal looks like”Every Proposal answers the same questions, so you can decide in one read.
| Field | Example Proposal |
|---|---|
| What | Turn on disk encryption on 6 MacBooks |
| Why | Your ISO 27001 blueprint requires disk encryption (A.8.1, A.8.24). |
| Evidence | Encryption is off on all 6. Two of them sent the most sensitive data to AI tools this month. |
| Impact | 6 devices, 6 people. Disruption: low. |
| When | Each laptop’s next idle window on power, over 3 days. |
| If you don’t | The A.8.1 gap appears in your audit evidence on your audit date. |
| Alternatives | Only the 2 most sensitive laptops first. |
| Rollback | One click; a new blueprint version without the change. |
Illustrative example. Proposals in your workspace use your devices, frameworks and dates.
Why AI-native
Section titled “Why AI-native”Most MDMs are a console of settings you push. Shadow MDM starts from the device and from your frameworks, and does the triage for you.
- It learns each laptop. Idle windows decide when a change lands, so people are not interrupted mid-call. Performance data sets the expected disruption.
- It knows what is sensitive. The same agent that masks data going to AI knows which laptops send the most sensitive data, and proposes tighter settings for those devices only (for example a shorter screen lock or stricter masking).
- It is proactive. It forecasts gaps before they bite: an audit date with failing controls, an OS version about to lose support, a laptop whose updates keep getting deferred, an AI tool spreading across devices with no policy.
- It never goes below the minimum. Models decide how, when and for whom. Whether a framework minimum applies is fixed by your frameworks, not by a model.
- It never acts alone. Engines and models only write Proposals; people approve them.
Enforcement Map
Section titled “Enforcement Map”One view of everything Shadow enforces: framework → control → rule → setting → devices → evidence. Each rule shows its confidentiality, integrity and availability impact, the controls that require it, who approved it, exceptions and their reasons, and its full history. Look at any past date, give your auditor a read-only role, and export signed evidence.
Compliance blueprints
Section titled “Compliance blueprints”Pick the frameworks you follow and Shadow chooses the blueprint: the minimum settings each framework needs, merged so the strictest wins.
| Framework | What it adds |
|---|---|
| ISO/IEC 27001 | Endpoint baseline with CIS Level 1 hardening; Level 2 optional (“Strict”) |
| ISO/IEC 42001 | AI blueprint: Shadow agent protected, AI traffic through Shadow, personal data masked, approved AI tools only, local model runtimes governed |
| SOC 2 | Endpoint baseline |
| PCI DSS v4.0.1 | Stricter passwords (12 characters), daily anti-malware definitions |
| HIPAA | 10-minute screen lock, encryption and remote-wipe readiness |
| GDPR | Endpoint baseline |
| EU AI Act | AI blueprint |
Blueprints in detail · Rule-to-control mapping
Works with the MDM you have
Section titled “Works with the MDM you have”If a device already reports another MDM, Shadow sends it no enforcement and the agent changes nothing. Posture is still reported, free, as evidence. You can keep Jamf, Intune or Kandji and still get the Enforcement Map and posture evidence; see Roll out Shadow with your MDM.
Security model
Section titled “Security model”- Blueprints are signed by Shadow Hub (Ed25519). Devices refuse unsigned or tampered blueprints.
- Hub is the source of truth; the device is treated as untrusted.
- Every applied, failed or rolled-back change becomes an activity entry with a signed receipt.
- Remote lock and wipe always need two different admins: one requests, another approves.
Pricing
Section titled “Pricing”Shadow MDM is included in Shadow Hub at no extra cost: posture checks, Proposals, approvals and staged rollouts. Hub is $3 per user per month, with a 30-day free trial and no card required.
See pricing · Shadow MDM vs traditional MDM
All Shadow MDM pages
Section titled “All Shadow MDM pages”- BlueprintsHow Shadow MDM builds a device blueprint from your frameworks: the endpoint baseline, the ISO 42001 AI blueprint, PCI DSS and HIPAA additions, CIS Level 1 vs Level 2.
- Enforcement MapThe Shadow MDM Enforcement Map traces framework to control to rule to setting to devices to evidence, with CIA impact, approver, exceptions, history and signed exports.
- Rule-to-control mappingEvery Shadow MDM rule with its CIA impact, platform settings and the ISO 27001, SOC 2, PCI DSS v4.0.1, HIPAA, ISO 42001, GDPR and EU AI Act controls it supports.
Frequently asked questions
What is Shadow MDM?
Device management, included in Shadow Hub, that manages laptop settings such as disk encryption, screen lock, firewall, updates and AI tool policy. It is built from scratch in Rust inside the Shadow agent you already run for AI privacy, so there is no second agent to install.
Does Shadow MDM change devices on its own?
No. On-device models and the Hub engine only write Proposals. Every change needs an admin's approval in Shadow Hub or Meridian, rolls out to a pilot group first and can be rolled back in one click. Remote lock and wipe need two different admins.
Do I need to replace Jamf, Intune or Kandji?
No. If a device reports another MDM, Shadow sends it no enforcement and the agent applies nothing; it only reports posture, which is free. You can keep your MDM and use Shadow for AI privacy, posture evidence and the Enforcement Map.
Which frameworks does it cover?
ISO/IEC 27001, ISO/IEC 42001, SOC 2, PCI DSS v4.0.1, HIPAA, GDPR and the EU AI Act. Shadow builds a blueprint from the frameworks you follow: the minimum each one requires, strictest setting wins. See Blueprints.
Can Shadow MDM lock or wipe a Mac remotely?
Not yet. macOS remote lock and wipe need the Apple MDM protocol, and Shadow's Apple MDM certificate is pending. Until then, macOS lock uses the Shadow agent's lock screen and the remote-wipe check reports as not ready. Windows lock and wipe work through the Windows MDM interface.
Does it enforce disk encryption on Linux?
No. On Linux, disk encryption (LUKS) is report-only: Shadow checks and reports it as evidence, but does not turn it on.
What does it cost?
Nothing extra. Shadow MDM, including posture checks and approving and rolling out changes, is included in Shadow Hub at $3 per user per month, with a 30-day free trial and no card required. See Pricing.
Is Shadow MDM certified for ISO 27001 or SOC 2?
Tools are not certified against these frameworks; organisations are. Shadow MDM helps you implement and evidence the endpoint controls those frameworks ask for, and shows which controls a device MDM does not cover.