New
Discovery Reference

How ThreatLynx discovers, scores, and governs connected applications

This reference explains the technical mechanisms behind ThreatLynx SaaS discovery — what data is collected, how risk is scored, and how governance workflows operate. Intended for security teams, IT administrators, compliance reviewers, and procurement due diligence.

Metadata-basedNo endpoint agentsRead-only API accessGoogle WorkspaceMicrosoft 365

ThreatLynx provides governance visibility and readiness support. Discovery is metadata and permission-based. It does not constitute compliance certification, legal assurance, or guaranteed detection coverage. Findings should be reviewed by qualified security, risk, and legal advisors.

How OAuth-based discovery works

ThreatLynx reads authorised application records from your workspace tenant APIs. No endpoint software is required and no workspace content is accessed.

OAuth grant discovery

ThreatLynx reads authorised OAuth client records from your workspace tenant APIs. Each record represents an application a user or administrator has granted access to — captured by the identity provider, not by endpoint software.

Google Workspace Admin SDK

For Google tenants, ThreatLynx calls the Workspace Admin SDK to enumerate domain-wide delegated authority and user-level OAuth tokens. This provides an inventory of authorised apps across your domain.

Microsoft Graph

For Microsoft 365 tenants, ThreatLynx reads enterprise application registrations, service principals, and delegated permission grants via Microsoft Graph.

Metadata-only analysis

Discovery is limited to OAuth metadata: app names, client IDs, granted scopes, user counts, vendor registration details, and governance status. ThreatLynx does not access the content of email, files, or messages.

Continuous or on-demand sync

Workspace sync can be triggered manually or scheduled. Each sync captures the current state of OAuth grants and compares against prior records to surface new or changed applications.

Single read-only connection

ThreatLynx connects using a service account or enterprise application with read-only access to OAuth/identity APIs. No agents are installed. No changes are made to your workspace during discovery.

Discovery data flow

A single read-only connection from ThreatLynx to your workspace admin APIs. No agent installation, no proxy, no traffic intercept.

Your workspace
Google Workspace
Microsoft 365Beta
Read-only APIOAuth metadata only
ThreatLynx
Discovery engine
Risk scoring
Governance console
Your team
Review & action
Security operations
IT administration
Compliance team

What ThreatLynx detects

Discovery captures OAuth metadata and governance signals — not content. The following data types are collected from workspace admin APIs.

Data typeDetail
Connected third-party applicationsAll apps with at least one active OAuth grant in the tenant.
OAuth scopes per applicationSpecific API permissions granted: mail read/write, files, calendar, directory, admin.
User authorisation countsNumber of users who have independently authorised each application.
Service principalsApplications acting as automation or integration identities without user context (M365).
AI-indicator applicationsApps matching known AI/LLM vendor patterns or names — flagged based on metadata, not content inspection.
MCP-connected systemsModel Context Protocol connected applications identified via OAuth patterns and vendor metadata.
Automation and workflow toolsApplications identified as workflow automation platforms (e.g. Zapier, Make) based on vendor metadata.
Governance statusApproved, Under Review, Restricted, or Blocked — as recorded in ThreatLynx by your team.
Risk score and bandComposite score 0–100 across four dimensions. Displayed per-application with contributing factors.
Vendor jurisdictionCountry of incorporation inferred from vendor registration data where available.

What ThreatLynx does not access

Understanding the boundaries of discovery is essential for security review and procurement due diligence.

Email content

ThreatLynx does not read, retrieve, or store the content of any email message, calendar event, chat message, or notification in your workspace.

File and document content

Documents, spreadsheets, presentations, and files stored in Google Drive, SharePoint, or OneDrive are not accessed. ThreatLynx records only that a connected app has a file-access scope — not what it accessed.

Employee behaviour or activity

ThreatLynx does not monitor employee behaviour, track individual user activity patterns, log keystrokes, or record screen interactions. It is a SaaS governance tool, not an employee monitoring product.

Endpoint data

No software is installed on endpoints. ThreatLynx does not access device configuration, locally installed applications, browser history, or network traffic from client devices.

Real-time event streams

Discovery reads point-in-time OAuth grant state from admin APIs. It does not tap into real-time event logs, webhook streams, or audit log feeds unless a specific integration is configured and consented to.

Credentials or tokens

ThreatLynx does not capture, store, or replicate OAuth access tokens, refresh tokens, or user passwords. It reads metadata records only.

Privacy architecture by design

The discovery API connection is scoped to the minimum permissions required to enumerate OAuth grant records. ThreatLynx does not request, store, or process any scope that would permit content access to email, files, or messages. Our Privacy Policy documents the full scope of data collected. Read the Privacy Policy →

Discovery sources

ThreatLynx currently supports two discovery sources. Additional sources are in development.

Google Workspace

Generally available
  • APIGoogle Workspace Admin SDK
  • ScopeDomain-wide OAuth clients + user-level tokens
  • Auth methodService account with domain delegation
  • Data returnedApp name, client ID, scopes, user count, grant date
  • Sync typeManual trigger or scheduled sync
  • Content accessNone — metadata only

Microsoft 365Beta

Feature coverage evolving
  • APIMicrosoft Graph v1.0
  • ScopeEnterprise apps, service principals, delegated grants
  • Auth methodApp registration in Entra ID (Azure AD)
  • Data returnedApp name, client ID, permissions, principal type
  • Sync typeManual trigger
  • Content accessNone — metadata only
Microsoft 365 discovery operates via Microsoft Graph. Scope depth and service principal discovery reflect the metadata Microsoft Graph exposes at the time of sync.

Risk scoring methodology

Every discovered application receives a composite risk score from 0–100. Scores are calculated from explicit, auditable factors — no black-box model.

Highest weight

OAuth scope sensitivity

Mail write/delete, directory admin, and delegated admin scopes carry the highest per-dimension scores. Profile-only or calendar read scopes score lower.

High weight

Scope breadth

An application granted access across multiple sensitive categories (mail + files + directory) scores higher than a single narrow-purpose scope.

High weight

User exposure count

Wider user authorisation amplifies risk. A high-sensitivity scope held by one user scores lower than the same scope granted across an entire department or domain-wide.

Medium weight

AI tool indicator

Applications matching AI/LLM vendor patterns receive additional weighting. AI services commonly process sensitive content to deliver features, raising data-exposure considerations.

Medium weight

Vendor jurisdiction

Applications registered outside Australia may raise data sovereignty and privacy compliance considerations — reflected as a contributing factor, not a disqualifying condition.

Modifier

Governance status

Unreviewed applications receive a higher score than those with a documented governance decision. Taking a governance action (approve, restrict, block) updates the active score.

Scores are indicative risk signals for governance prioritisation. They are not definitive security assessments and should not be the sole basis for enforcement decisions.

Agent governance discovery

ThreatLynx identifies likely non-human identities and AI-connected applications using rules-based heuristics applied to OAuth metadata. Detection is confidence-based, not absolute.

OAuth AI agents

Applications identified as autonomous or AI-powered via name matching against known vendor patterns and OAuth scope analysis. Flagged for governance review as higher-sensitivity identities.

Service principals

Non-human identities created for application-to-application access in Microsoft 365 environments. Discovered via Microsoft Graph enterprise application records.

Workflow automations

Applications associated with workflow automation platforms that act across multiple users and data sources — identified via vendor metadata and OAuth scope breadth.

MCP-connected systems

Model Context Protocol integrations identified via vendor patterns in OAuth client metadata. Detection is confidence-based and should be reviewed by application owners.

AI copilots and extensions

Assistant-type integrations (chat copilots, writing assistants, browser extensions) with workspace data access. Identified from OAuth grant metadata and vendor classification.

Autonomous integrations

Applications exhibiting automation-style OAuth patterns: broad scopes, service-level grants, or persistent access without a single user owner. Flagged as requiring governance review.

Agent detection confidence model
High confidence

Known AI/agent vendor + sensitive scope access (mail, files, directory). Both vendor classification and scope classification present.

Medium confidence

Partial signals — vendor name pattern only, or scope-only match without confirmed AI vendor classification.

Low confidence

Weak or single-signal match. Governance review recommended before any classification decision.

Agent identity signals are governance readiness support — not a certified AI inventory, legal attestation, or regulatory determination. Findings should be reviewed by security, risk, or application owners before formal use.

Compliance framework mapping

ThreatLynx discovery data supports governance readiness for Australian and international frameworks. Coverage is partial — not certification.

ASD Essential Eight

Direct

Application Control (ML2/ML3) · User Application Hardening

ThreatLynx discovery provides an inventory of third-party OAuth-connected applications across your tenant — directly relevant to Application Control maturity and the assessment of authorised vs. unauthorised software. User Application Hardening is supported by identifying which applications have been granted elevated scopes by individual users.

APRA CPS 234

Supports

Information Asset Register · Third-party Assessment

CPS 234-regulated entities are expected to maintain an information asset register and identify risks from third-party information security capability. ThreatLynx provides a continuous SaaS inventory with governance decisions that can be used as evidence in CPS 234 information security capability assessments.

Australian Privacy Principles (APP 11)

Supports

Security of Personal Information

APP 11 requires organisations to take reasonable steps to protect personal information from misuse, interference, and unauthorised access. ThreatLynx helps identify which third-party applications have been granted OAuth access to personal information held in Workspace environments.

ISO/IEC 27001:2022

Partial

A.8 Asset Management · A.9 Access Control

ThreatLynx supports asset inventory (Annex A.8) and access control review (Annex A.9) by maintaining a discoverable record of connected SaaS applications and their granted permissions. This does not constitute ISO certification or a complete implementation of Annex A controls.

Regulatory references are informational. ThreatLynx does not provide legal advice, compliance certification, or guaranteed compliance outcomes. Coverage designations reflect the nature of support provided, not audit attestation. Consult qualified legal and compliance counsel for specific regulatory obligations.

Limitations and confidence

Effective governance begins with understanding what the tool can and cannot do. These limitations apply to all ThreatLynx deployments.

01

API access is required

Discovery depends on administrative API access to your Google Workspace or Microsoft 365 tenant. If the required Admin SDK or Graph API permissions are not granted, some or all applications may not appear in the inventory.

02

Microsoft 365 discovery depends on Graph metadata

Microsoft 365 discovery via Microsoft Graph depends on the permission scopes and metadata returned by the tenant. Service principal discovery and permission scope depth reflect what Microsoft Graph exposes at the time of sync.

03

Metadata-based confidence

Application classification (AI, automation, MCP) is based on vendor name patterns and OAuth metadata — not content inspection. Some applications may be misclassified, and some categories of interest may not be represented in current detection rules.

04

Not all OAuth clients are visible

Certain application types — particularly those using OAuth flows not recorded in standard admin APIs, or applications approved by individual users before ThreatLynx was connected — may not appear in the initial inventory.

05

Risk scores are indicative

Scores reflect risk prioritisation for governance review, not a definitive security assessment. They should be used as a starting point for human review, not as the sole basis for enforcement decisions.

06

Governance action is manual

ThreatLynx records governance decisions but does not automatically revoke OAuth tokens or block application access. Revocation must be performed by an administrator through the workspace admin console.

ThreatLynx provides governance readiness support for Australian organisations using Google Workspace and Microsoft 365. Discovery coverage, score accuracy, and governance outcomes depend on accurate and complete administrative API access. All findings should be reviewed by qualified security, IT, and legal professionals before use in formal compliance assessments or enforcement decisions.

See discovery on your workspace data

Request a guided walkthrough to explore ThreatLynx SaaS discovery, risk scoring, and governance workflows in your Google Workspace or Microsoft 365 environment.