Anthropic (Claude)

Anthropic (Claude) is an AI assistant platform that provides enterprise administration capabilities for API key inventory, user and group management, and compliance activity monitoring.

The Anthropic (Claude) adapter connects Axonius to the Anthropic API to provide visibility into your organization's Anthropic usage: who has access, which groups they belong to, and which API keys are active. It supports three distinct API modes - Enterprise, Compliance and Enterprise Analytics - each backed by a different API key type and fetching a different set of entities.

Use Cases the Adapter Solves

API Key Inventory and Lifecycle Management (Enterprise only)

  • What it solves: API keys are a primary attack surface for AI services. Leaked, expired, or orphaned keys pose a significant risk of unauthorized access to LLM capabilities and, indirectly, to data processed through Anthropic.

    Security value:

    • Provides a complete inventory of all API keys issued within the organization.
    • Surfaces key metadata including creation date, expiration date, current status (active / inactive), and type.
    • Enables security teams to detect keys that are expired but still present, keys with no expiration date (indefinite lifetime), and keys that may have been created outside of normal provisioning processes.

    User Access Visibility

    What it solves: Understanding who has access to Anthropic β€” both at the organization level and at the workspace level β€” is essential for access reviews, off-boarding validation, and least-privilege enforcement.

    Security value:

    • Provides a complete roster of users with access to the organization's Anthropic account.
    • Captures user roles (organization_role / role), enabling differentiation between admins and standard users.
    • Records when users were created (created_at) and when they were added to a workspace or group (added_at), supporting audit trail and onboarding/off-boarding reviews.
    • Both API modes (Enterprise and Compliance) produce user entities, allowing the mode to be selected based on available credential type without losing user coverage.

    Workspace and Group Access Control Auditing

    What it solves: Anthropic organizes resources into workspaces (Enterprise) and groups (Compliance). Knowing which users belong to which workspaces/groups is critical for access boundary enforcement and blast-radius analysis.

    Security value:

    • Maps the full membership of every workspace or group in the organization.
    • Enables detection of over-privileged groups, unexpected cross-team membership, and groups with no members (zombie groups).
    • Captures group descriptions, source types, and assigned roles β€” providing context for who owns the group and what it controls.
    • Supports Axonius policy enforcement rules that flag users with access to multiple high-privilege workspaces.

Asset Types Fetched

  • Secrets (API Keys) - Enterprise mode only , Users, Roles, Groups (Workspaces or Compliance Groups), Knowledge Sources, AI Agents, AI Models, AI Tools

Data Retrieved through the Adapter

Secrets (API Keys) - Enterprise mode only - Fields such as Name, Created Time, Expires At, Status

Users - Fields such as Email, Username, Display Name, User Created Date

Groups (Workspaces or Compliance Groups) - Fields such as Name, Display Name, Description, Source Type

Roles- Fields such as Name, Description, Created Time, Permissions

Users (Enterprise Analytics mode) - Fields such as Analytics User ID, Analytics Last Activity Date, Chat Message Count

AI Models - Fields such as Name, Version, Family, Provider, Status, Supported Modalities, Model Max Input Tokens, Model Max Output Tokens, Model Fallback, and model capabilities (for example Supports Image Input, Supports PDF Input, and Supports Thinking)

Before You Begin

Required Ports

  • TCP port 443 (HTTPS)

Authentication Methods

The Anthropic adapter supports the following authentication methods based on API key type:

  • Enterprise API Key (Admin API) - Uses an Admin API key (starting with sk-ant-admin01-...)
  • Compliance API Key - Uses a Compliance Access Key (starting with sk-ant-api01-...)
  • Enterprise Analytics API Key - Uses an Analytics API key with the read:analytics scope

APIs

Axonius uses the Claude API (version 2023-06-01). The following endpoints are called:

Enterprise Mode Endpoints:

  • GET /v1/organizations/api_keys - Retrieves all API keys in the organization
  • GET /v1/organizations/users - Retrieves all users in the organization
  • GET /v1/organizations/workspaces - Retrieves all workspaces
  • GET /v1/organizations/workspaces/{workspace_id}/members - Retrieves members of each workspace
  • GET /v1/files - Retrieves the knowledge source files of each workspace (when Fetch Knowledge Sources is enabled)

Compliance Mode Endpoints:

  • GET /v1/compliance/organizations - Retrieves organization UUID (helper endpoint)
  • GET /v1/compliance/organizations/{org_uuid}/users - Retrieves all organization users
  • GET /v1/compliance/groups - Retrieves all compliance groups
  • GET /v1/compliance/groups/{group_id}/members - Retrieves members of each group
  • GET /v1/compliance/apps/projects - Retrieves projects (when Fetch Knowledge Sources is enabled)
  • GET /v1/compliance/apps/projects/{project_id}/attachments - Retrieves the attachments of each project (when Fetch Knowledge Sources is enabled)

Enterprise Analytics Mode Endpoints:

  • GET /v1/organizations/analytics/users - Retrieves the users with Claude activity
  • GET /v1/organizations/analytics/user_cost_report - Retrieves the cost of each user
  • GET /v1/organizations/analytics/user_usage_report - Retrieves the token usage of each user

Endpoints for AI Agents (Enterprise and Compliance modes):

  • GET /v1/agents - Retrieves AI agents (when Fetch AI Agents is enabled)
  • GET /v1/deployments - Retrieves the deployments of each AI agent (when Enrich AI Agents with Deployments is enabled)
  • GET /v1/environments - Retrieves the environment of each deployment (when Enrich Deployments with Environments is enabled)
  • GET /v1/agents - Retrieves the tools configured on each AI agent (when Fetch Tools is enabled)

Endpoints for AI Models (Enterprise mode):

  • GET /v1/models - Retrieves the AI models available to the organization (when Fetch Models is enabled)

Required Permissions

Enterprise Mode - Organization Admin API Key

To use Enterprise mode, you need an Organization-level Admin API key with the following scopes:

  • read:organization:api_keys / admin role
  • read:organization:users / admin role
  • read:organization:workspaces / admin role
  • read:organization:workspaces / admin role

Compliance Mode - Compliance API Key

To use Compliance mode, you need a Compliance API key with the following scopes:

  • read:compliance:organizations / Parent Org Compliance
  • read:compliance:users / Parent Org Compliance Access Key
  • read:compliance:roles / Parent Org Compliance Access Key
  • read:compliance:permissions / Parent Org Compliance Access Key
  • read:compliance:groups / Parent Org Compliance Access Key
  • read:compliance:groups / Parent Org Compliance Access Key

Enterprise Analytics Mode - Analytics API Key

To use Enterprise Analytics mode, you need an Analytics API key with the read:analytics scope. Only the organization's primary owner can create this key, in Claude under Organization settings > API. Admin and Compliance API keys cannot access the Enterprise Analytics API. Cost and token usage data is returned only for usage-based Enterprise plans.

Fetching Knowledge Sources

To use the Fetch Knowledge Sources advanced setting, the API key must have the following access:

  • Enterprise mode - A personal or service account API key of an organization admin that is not scoped to a single workspace. Admin API keys cannot access files.
  • Compliance mode - A Compliance API key with the read:compliance_user_data scope.

Supported From Version

Supported from Axonius version 9.0

Connecting the Adapter in Axonius

Navigate to the Adapters page, search for Anthropic, and click on the adapter tile.

Click Add Connection.

To connect the adapter in Axonius, provide the following parameters:

Required Parameters

  1. Domain - Base URL for the Claude API. Defaults to https://api.anthropic.com. Change only when routing through a proxy.
  2. API Type (default: compliance) - Selects the mode and determines which entities are fetched and which endpoints are called: Enterprise (to fetch Secrets, Users, and Groups), Compliance (to fetch Users and Groups, but does not expose API keys), or Enterprise Analytics (to fetch Users with Claude activity, cost, and token-usage analytics).
    The adapter automatically enables the appropriate endpoints based on the selected API Type (Enterprise, Compliance, or Enterprise Analytics) connection setting. If you want to fetch all API types, you need a separate adapter connection for each one.
  3. API Key - The secret key for the selected API type. This is passed as the x-api-key request header. The keys are not interchangeable.
  • API Key (required) - An Admin API key (starting with sk-ant-admin...) from the Claude Console.


Optional Parameters


  1. Verify SSL - Select whether to verify the SSL certificate of the server against the CA database inside of Axonius. For more details, see SSL Trust & CA Settings.
  2. HTTPS Proxy - Connect the adapter to a proxy instead of directly connecting it to the domain.
  3. HTTPS Proxy User Name - The user name to use when connecting to the value supplied in Domain via the value supplied in HTTPS Proxy.
  4. HTTPS Proxy Password - The password to use when connecting to the server using the HTTPS Proxy.

To learn more about common adapter connection parameters and buttons, see Adding a New Adapter Connection.

Advanced Settings

Note:

  • Global Endpoints Config - Click on > to open the following settings:
    • Fetch Knowledge Sources (Enterprise: Files / Compliance: Projects Knowledge) (default: false) - Select this option to fetch knowledge sources: files in Enterprise mode, or project knowledge in Compliance mode. (available only to customers with Security for AI.)
    • Fetch AI Agents (default: false) - Select this option to fetch AI agents. (available only to customers with Security for AI.)
    • Enrich AI Agents with Deployments (default: false) - Select this option to enrich AI agents with their deployments. To use this setting, select Fetch AI Agents. (available only to customers with Security for AI.)
    • Enrich Deployments with Environments (default: false) - Select this option to enrich deployments with their environments. To use this setting, select Enrich AI Agents with Deployments. (available only to customers with Security for AI.)
    • Fetch Tools (default: false) - Select this option to fetch AI Tools from Anthropic agents. Each tool configured on an agent becomes an AI Tools asset. (available only to customers with Security for AI.)
    • Fetch Models (default: false) - Select this option to fetch AI models. Each model becomes an AI Models asset. Enterprise API type only, with a personal or service-account API key that is not scoped to a workspace. (available only to customers with Security for AI.)
    • Analytics User Activity Lookback Days (default: 30) - Enterprise Analytics API type only. Enter the number of days of Claude user activity to fetch, from 1 to 365. Users with no activity in this period are not fetched.
    • Analytics Cost And Token Usage Lookback Days (default: 30) - Enterprise Analytics API type only. Enter the number of days of per-user cost and token usage to fetch, from 1 to 31. Cost and token usage are added only to users fetched by Analytics User Activity Lookback Days, so keep this value at or below it.

Did this page help you?