Frequently asked questions
Everything you need to know in order to get started and be successful with ARGOS.
Getting started
How do I sign up to ARGOS?
Browse to app.argos-security.io/register, log in with your Microsoft Entra ID or Google account, add your details and click sign up. No credit card is required.
The account you log in with must have a valid email address.
Why is Entra ID consent required to log in?
Depending on your company's Entra ID configuration, the first login with a Microsoft Entra ID account may ask for admin consent. That is a normal step when logging in to any SaaS for the first time, and it is controlled by your tenant, not by ARGOS.
If the login screen offers it, request admin consent and an Entra ID administrator can grant it. Otherwise an administrator can grant it directly by opening https://login.microsoftonline.com/{entra-tenant-id}/adminconsent?client_id=2f1d30d4-feb5-4743-879a-f192ef3eeeb4, replacing {entra-tenant-id} with your Entra ID tenant id.
What does an ARGOS assessment cover?
Microsoft Azure, Microsoft Entra ID, Microsoft 365 (Exchange Online, SharePoint, Teams and Intune), AWS and Azure AI Foundry. ARGOS checks configuration, identities and permissions, tests reachability for supported services, connects what it finds into attack paths and diagrams, and produces a customer-focused report.
See how a cloud security assessment works and the Entra ID security assessment.
Does ARGOS support single sign-on?
Yes. ARGOS logs you in through your existing identity provider, Microsoft Entra ID or Google, and stores no passwords of its own.
Does ARGOS support multi-factor authentication?
Yes. MFA is inherited from the identity provider you log in with: if MFA is configured on that account, ARGOS prompts for it too.
We recommend MFA, ideally with a FIDO key, on every account that accesses ARGOS.
Do MFA and SSO cost extra?
No. SSO and MFA are included on every plan, and we intend to stay off the SSO Wall of Shame.
Which roles can I give my team?
Four: Reader, Contributor, Cloud Admin and Account Admin.
Reader, Contributor and Cloud Admin see only the cloud connections assigned to them, which is how a service provider keeps each engineer inside their own customers. Account Admin sees every connection in the account.
Connecting your clouds
What do I have to deploy to use ARGOS?
Almost nothing. ARGOS is agentless and works through read-only, management-plane API access.
- One-off scan: read-only access to Azure, Entra ID or AWS, using your own sign-in.
- Continuous scans: an Entra ID app registration (for Azure, Entra ID and Microsoft 365) or an IAM role (for AWS) that ARGOS can use.
No agents, no infrastructure in your cloud, and no dependency on services such as Microsoft Defender for Cloud being enabled, configured or paid for.
Does ARGOS change anything in my cloud?
No. Assessment is agentless and read-only: ARGOS reads configuration through the clouds' management APIs and never reads data-plane content such as files, blobs, database rows or object contents.
The one exception is optional. Where remediation is enabled on your plan, you can apply selected fixes from ARGOS, and that needs write permissions you grant separately for the resources being changed. Without them, ARGOS cannot change anything.
How do I onboard Azure?
One-off scan: Reader on the subscriptions you want to scan is enough. Sign in with Entra ID from the dashboard, or paste an access token, and ARGOS creates a temporary connection for that scan. If you are a guest in the tenant, or user permissions are restricted, an Entra ID administrator may need to grant access; the dashboard explains the steps.
Continuous scans: create an app registration with a client secret, a certificate or a federated credential, then assign it Reader on each subscription you want monitored. Reader at management-group scope is optional and adds Azure Policy governance and PIM visibility. The Organization.Read.All Graph permission is optional; it is only used to show the tenant's display name.
Remediation, where enabled on your plan, needs the matching write permissions on the subscription.
How do I onboard AWS?
ARGOS assumes a cross-account IAM role in each account, with an external ID generated by ARGOS, so no AWS secrets are stored. Your My Account page holds the ARGOS account ID, the external ID and a CloudFormation template that creates the assessment role; a second template handles AWS Organizations discovery, and StackSets roll the role out to many accounts at once. Terraform is not offered.
The role needs the managed policies SecurityAudit, ViewOnlyAccess and AWSLambda_ReadOnlyAccess (used to check Lambda function code for secrets), plus a small custom policy allowing backup:GetBackupPlan, backup:ListBackupPlans, iam:ListAccountAliases, macie2:GetMacieSession, macie2:ListClassificationScopes and macie2:GetClassificationScope.
Then add the account in ARGOS with its account ID, the role name and the regions to monitor. The same policies apply to a one-off scan; the videos page has a short walkthrough. Remediation, where enabled on your plan, needs write access for the services being remediated.
How do I onboard Entra ID and Microsoft 365?
All access is read-only, and there are two ways in.
One-off scan (delegated sign-in). Sign in with Entra ID from the dashboard, or paste a Microsoft Graph access token. The user should hold the Global Reader role and, for Exchange Online, the View-Only Organization Management role. ARGOS uses only the delegated scopes listed below.
Continuous scans (app registration). Create an app registration, grant it the application permissions listed below with admin consent, and add its client ID and client secret in the dashboard. Sites.Read.All is only needed if you opt in to the SharePoint assessment, and the PIM permissions are optional.
Full permission list
Delegated scopes for a one-off scan. With the Microsoft Graph PowerShell module, three lines produce the token:
Connect-MgGraph -Scopes "Application.Read.All", "Directory.AccessAsUser.All", "Directory.Read.All", "Domain.Read.All", "openid", "profile", "User.Read.All", "UserAuthenticationMethod.Read.All", "Policy.Read.All", "AuditLog.Read.All", "EntitlementManagement.Read.All", "Synchronization.Read.All", "IdentityRiskEvent.Read.All", "OnPremDirectorySynchronization.Read.All", "RecordsManagement.Read.All", "RoleAssignmentSchedule.Read.Directory", "RoleEligibilitySchedule.Read.Directory", "RoleManagementPolicy.Read.Directory", "SharePointTenantSettings.Read.All", "TeamworkAppSettings.Read.All", "SecurityEvents.Read.All", "DeviceManagementConfiguration.Read.All", "DeviceManagementManagedDevices.Read.All", "DeviceManagementApps.Read.All", "DeviceManagementServiceConfig.Read.All", "ThreatIntelligence.Read.All", "RoleManagement.Read.Directory", "SecurityIncident.Read.All", "Device.Read.All"
$data = Invoke-MgGraphRequest -Uri "https://graph.microsoft.com/v1.0/me" -Method GET -OutputType HttpResponseMessage
$data.RequestMessage.Headers.Authorization.Parameter
Application permissions for continuous scans:
- Application.Read.All
- AuditLog.Read.All
- DeviceManagementConfiguration.Read.All
- DeviceManagementManagedDevices.Read.All
- DeviceManagementApps.Read.All
- DeviceManagementServiceConfig.Read.All
- Device.Read.All
- Directory.Read.All
- Domain.Read.All
- EntitlementManagement.Read.All
- IdentityRiskEvent.Read.All
- OnPremDirectorySynchronization.Read.All
- Organization.Read.All
- Policy.Read.All
- RecordsManagement.Read.All
- RoleAssignmentSchedule.Read.Directory
- RoleEligibilitySchedule.Read.Directory
- RoleManagement.Read.Directory
- RoleManagementPolicy.Read.Directory
- SharePointTenantSettings.Read.All
- SecurityEvents.Read.All
- SecurityIncident.Read.All
- Synchronization.Read.All
- TeamworkAppSettings.Read.All
- ThreatIntelligence.Read.All
- User.Read.All
- UserAuthenticationMethod.Read.All
How much does it cost me to run ARGOS?
Besides the subscription there is no other cost associated with running ARGOS against Azure.
On AWS the extra API calls can add to your AWS CloudTrail cost; contact [email protected] to talk it through.
Using ARGOS remediation to change a resource can change what that resource costs.
Scans and policy sets
One-off or continuous: which do I need?
A one-off scan is a single assessment: you authenticate with your own sign-in, ARGOS scans the environment, and you get the report. It suits an engagement, a pre-sales assessment or a point-in-time review.
A continuous scan runs every day against a connected tenant or account, keeps findings current, and feeds notifications and integrations. Connect an environment for continuous scanning when you want to track it over time rather than assess it once.
How often do continuous scans run?
Once a day, for every connected tenant and account.
While a scan runs, progress is shown per connection in the dashboard, so you can see which environment is still being collected and which is done.
Can I choose which rules a scan runs?
Yes, with scan policy sets. Pick a built-in set for a connection: All ARGOS Rules, Attack Paths Only, or one per supported framework and cloud (ISO 27001 for Azure or Essential Eight for Entra ID, for example), or define your own set from the rule catalogue.
No selection means all rules run.
Can we ignore resources from an ARGOS scan?
For selected resource types, yes. Tagging a resource argos-ignore = true excludes it from continuous scans. The tag is honoured by a subset of Azure rules and some AWS collectors, not by every rule, so check the result after the next scan.
How to tag resources: Azure, AWS. To exclude a single finding regardless of resource type, ignore the detection instead (next question).
Can we ignore a detection?
Yes. Open the detection and click Ignore; it leaves graphs and scoring until you find it again among the ignored detections and click Unignore.
Both actions are recorded in the audit log with the user and the reason given.
Results and reports
What reports will I get?
When a scan completes, ARGOS emails you a notice with a link to the dashboard. The report itself is downloaded there, by any user authorised for that connection, as a customer-focused assessment report in a single ZIP. The file is not attached to the email.
It covers the executive summary, priority risks, attack path examples with their confidence, a recommended remediation plan and timeline, compliance by environment, Conditional Access gaps, Zero Trust, agentic AI and Copilot Studio, internet exposure, the Azure network and architecture summary, and Azure service retirements. Report files are kept for 7 days after generation. See the sample report.
Does ARGOS create cloud architecture diagrams?
Yes, three kinds, all generated from the scan.
- The architecture view: a generated diagram of the assessed environment.
- The unified graph on the Summary page, which can show, for example, all Azure VMs with a public IP and the resources around them, or every virtual network in the environment.
- A detection diagram for each finding, showing the immediate environment around the affected resource.
Across all of them ARGOS shows lateral movement paths through the network and cloud identities, security and compliance issues, and the relationships between resources that cloud-native tools hide at scale.
What does "confirmed" mean on an attack path?
Every step of an attack path carries its own proof, and the path's confidence is computed from those steps, never asserted by a rule. Confirmed means every step is proven from collected configuration. Conditional means a step depends on something ARGOS cannot observe, and that assumption is stated on the path. Indeterminate means the evidence was incomplete.
Confidence is shown on every path in the dashboard and in the report.
What is the difference between a security failure and an architecture review?
A security failure is a misconfiguration ARGOS can call wrong on its own: it counts against your security score and fails the benchmark controls it maps to.
An architecture review is a design choice whose risk depends on context ARGOS cannot see. Reviews are listed separately and never count against the score, but they still fail an applicable benchmark control, so a compliance view stays honest.
How can I use ARGOS AI?
Open any detection at app.argos-security.io/detections, wait for the diagram to load and click the ARGOS AI button ("Explain this to me"). ARGOS AI explains the finding in context, including attack paths and remediation advice.
The same engine writes the report's executive summary and recommendations, the network and agentic AI summaries, choke-point fix guidance and the Conditional Access gap remediation plan. How it handles your data is covered under Data and security.
Can I work with ARGOS via API?
Yes. Every ARGOS user gets their own API key, and the developer portal at dev.argos-security.io holds the API definitions, so day-to-day actions can be scripted from curl, bash, PowerShell or Postman.
Endpoints include starting a one-off scan of Azure, AWS or Entra ID, reading all or single detections and inventory, and adding or removing Azure subscriptions, Entra ID tenants and AWS accounts from continuous scanning.
How do I connect an AI agent to ARGOS with MCP?
The ARGOS MCP server lets an AI agent such as Claude, ChatGPT or Copilot query your assessment data: findings, attack paths, remediation priorities, compliance results and inventory. Access is read-only, so an agent cannot change anything in ARGOS or in your cloud.
An Account Admin creates an integration under Settings, Integrations, AI Agents (MCP) at app.argos-security.io/account/ai-agents, choosing the tools and connections the agent may use. Then add https://mcp.argos-security.io as an MCP server in your agent.
- Signed-in users: people sign in with their normal ARGOS login, and the agent sees only what both they and the integration may see.
- Client credentials: for unattended agents and scripts. ARGOS shows the secret once; rotate or remove it at any time.
AI agent access is enabled per account and is not available on trial accounts; contact [email protected] to turn it on.
I already have a SIEM (Microsoft Sentinel, Splunk, ...).
Keep it. ARGOS complements a SIEM rather than replacing it.
When an incident lands in your SIEM, open the ARGOS diagram for the resource involved: it shows the attack paths, lateral movement paths and related resources and their issues, so you can see what else may be affected without researching it by hand.
I already have a CSPM.
Keep it. A CSPM gives you the list of issues; ARGOS helps you decide which ones matter.
Open the ARGOS diagram for a resource your CSPM flagged and you see its attack paths, lateral movement paths and related resources and their issues, which is the context that decides the priority.
Integrations
How do I send notifications to Slack?
ARGOS can post new detections into a Slack channel you own. All it needs is an incoming webhook to that channel.
Create it by clicking the "Add to Slack" button on app.argos-security.io/account/notifications.
How do I send notifications to Microsoft Teams?
ARGOS can post new detections into a Teams channel you own, through either a Teams incoming webhook or a Teams Workflow.
Create the webhook by following Microsoft's steps, then add its URI on the Notifications tab at app.argos-security.io/account/notifications.
How do I create issues in Atlassian Jira?
On app.argos-security.io/account/jira, add your Jira account's email address, an API token, your workspace URI (in the form https://yourworkspace.atlassian.net/), the default project and the issue type ARGOS should create.
Once configured, every detection at app.argos-security.io/detections shows an "Export to Jira" button. The export can also be automated, so each new detection is added to Jira as ARGOS finds it.
How do I create incidents in ServiceNow?
On app.argos-security.io/account/service-now, add your ServiceNow instance name and the user name and password of a ServiceNow user allowed to create incidents.
Once configured, every detection at app.argos-security.io/detections shows an "Export to ServiceNow" button. ServiceNow export is on demand from a detection; it is not an automatic notification.
How do I send events to Microsoft Sentinel?
Browse to app.argos-security.io/account/sentinel and paste your Sentinel workspace ID and primary key.
New detections are then sent to Sentinel automatically and can be queried from the ARGOS_CL table. A Boomi integration is available on request.
Data and security
Where does ARGOS store my data, and is it secure?
You choose the region when you sign up: Australia (Australia East) or the European Union (North Europe). ARGOS is hosted on Microsoft Azure; the components that scan AWS run in AWS.
Cloud credentials are encrypted with AES-256-GCM using a key held in Azure Key Vault, ARGOS stores no passwords because every account logs in through SSO, and access to your clouds is read-only. The Trust Centre has the detail on data handling, encryption and how ARGOS reaches your cloud.
How long is my data kept, and what happens when we cancel?
Your data is retained for as long as you have an active ARGOS subscription, and you stay in control of it: removing a cloud connection runs a staged, verified purge of that connection's inventory, findings, graphs, snapshots and reports. Report files are kept for 7 days after generation.
Customer data is deleted promptly after the subscription ends. See data handling in the Trust Centre.
How does ARGOS AI work, and does it use my data to train models?
ARGOS AI runs on Azure OpenAI over private endpoints, deployed in the EU (Sweden Central) for every customer region, including Australia. It is sent finding context and configuration, never credentials.
Customer data is never used to train any model, under Azure OpenAI's data handling terms. See AI in the Trust Centre.
No questions match that search.