AzureHound
Use this skill for AzureHound / Entra ID BloodHound query design and attack-path triage.
Required context
- Authorized Azure tenants/subscriptions and whether AzureHound/Entra collection is present.
- Known object IDs, tenant names, privileged roles, subscriptions, or resource scopes.
- Hybrid collection availability when paths cross AD, GitHub, Okta, or SCIM.
Workflow
- Read
../../references/docs/bloodhound-query-methodology.mdand../../references/docs/azurehound-methodology.md. - Search
../../references/query-indexes/azurehound.mdfor a matching Query Library pattern. - Inspect the snapshot and confirm
AZ*labels/edges before adapting. - Use exact
objectidfilters when possible and bound broad tenant paths. - Explain each path segment by platform and collector source.
Common pivots
- Users/groups/service principals to privileged Entra roles.
- App owners, app role assignments, credentials, and Graph API permission edges.
- Managed identities to Azure resources.
- Subscription/resource group/VM/Key Vault control paths.
- AAD/Entra Connect and synced identity bridges.
- GitHub/OIDC or Okta/SCIM hybrid paths when data is present.
Output
Use the shared output contract from $bloodhound-query and include Azure-specific caveats such as display-name ambiguity, tenant-scale query cost, non-traversable Graph API edges, and collector freshness.