35 · 123
Collectors run · distinct findings they can produce. Public records only.
METHODOLOGY
SCOUTz records what it checked, when it checked it, which source supported the result, and whether missing coverage limits the conclusion.
EVIDENCE · INSIGHT · PROOF
Published domain, DNS, email authentication, certificates and external evidence.
Supported Microsoft 365 identity, app, permission, licensing and configuration evidence.
No message bodies, private files, chats or prompts. Unread is not clean.

PUBLISH THE UNIT
Domain and cloud are different access layers. The platform total also includes fusion and correlation passes that run on top of collector output.
Collectors run · distinct findings they can produce. Public records only.
Collectors run · distinct findings they can produce, including the public-domain roster.
Distinct finding types across 91 emitting modules and correlation passes.
All 601 finding types carry an authored explanation, consequence, recommendation, and the evidence they were drawn from. The build fails if one is missing.
538 registered check IDs keep every finding defensible. The ledger separates 440 scored findings, 107 context observations, and 22 declared non-reads. A non-read is never a clean result.
INDEPENDENT DISCOVERY
PSA, RMM, and specialist tools remain valuable. SCOUTz discovers, observes, normalizes, correlates, and asks what the combined evidence actually supports.
ASSESSMENT BOUNDARY
Specialist reviews are not dishonest; they were designed for a narrower job. SCOUTz begins with the company and leaves the implementation choice with the MSP.
SPECIALIZED VENDOR ASSESSMENT
The vendor's specialty
Evidence in that boundary
Needs related to that capability
The vendor's implementation
SCOUTz CLIENT ASSESSMENT
Start with the organization
Gather supported evidence
Name what needs attention
Define what better means
Product, service, process, project—or nothing
EVIDENCE STATES
These are distinct evidence states, not a mathematically linear score ladder.
Directly supported or completed validation
Seen in the named source at the recorded time
Supported, but not directly proven
Plausible and still needs investigation
The source cannot establish the state
The source could not be read
REAL APP CAPTURE · CLIENT DETAILS REMOVED

A finding describes the condition, evidence, affected object, unknowns, and recommended outcome—not an automatic instruction to buy Vendor X. Any future vendor recommendation must be explicit, transparent, clearly separate from the underlying finding, and unable to change what the evidence says.
Vendor assessment or client assessment? →Directly supported by a named authoritative source or completed validation.
Seen by SCOUTz from the disclosed source at the recorded time.
Supported by evidence, but not directly proven by the available source.
A plausible condition or opportunity that still requires investigation.
The available evidence cannot establish the state. Unknown is not clean.
The source could not be read because of availability, format, access, or permission.
ASSESSED · PARTIAL · NOT_ASSESSED · PERMISSION_LIMITED · LICENSE_LIMITED. A permission- or license-limited source never becomes a clean score by omission.
DOMAIN BOUNDARY
DNS, email authentication, HTTPS, certificates, headers, subdomains, exposed services, and related external evidence. No proof of internal controls, recovery, people, or process.
MICROSOFT 365 BOUNDARY
Supported identity, application, security, and managed-device context. No mail, files, chats, documents, prompts, or automatic changes.
A public observation and an authorized tenant record can support different kinds of claims. Customer authorization can strengthen the available source, but it never removes coverage, permission, license, provider, or readability limits.
Why permission follows the question →Start with the business or technical question instead of the broadest API permission available.
Use the least intrusive source that can responsibly support the answer.
If deeper evidence is necessary, state the purpose and boundary before collection begins.
Every image labeled Real Product is captured from the working application and anonymized. Every designed explanation is labeled Illustrative. We do not use AI-generated interfaces as evidence that a feature exists.
Read the authenticity standard →Bring dated evidence, clear limits, and useful questions.