Threat hunts for Azure

48 hunts covering Azure, each with a hypothesis, the queries that test it and what the hunt cannot see.

48 hunts, newest first

All 486 hunts
  1. Agent-Framework Exfiltration to Webhook.site

    An intruder is exfiltrating harvested cloud and developer credentials from Python runtimes to a malicious webhook.site endpoint following the agent-framework PyPI compromise.

    High Elastic Security Labs Part 2 of 2 · 3 queries, 8 steps credential access, execution
  2. Malicious Python Package Persistence and Secret Theft

    An intruder has installed a compromised version of the agent-framework package which persists via a .pth hook and harvests cloud credentials by reading environment variables from the /proc filesystem and local configuration files.

    High Elastic Security Labs Part 1 of 2 · 4 queries, 11 steps credential access, execution
  3. High-Tempo Host Execution and C2 Burst Analysis

    An adversary is using agentic offensive frameworks like ARTEX or CyberStrikeAI to automate initial access and command execution, evidenced by high-frequency network bursts and browser processes spawning system shells.

    High Elastic Security Labs Part 2 of 2 · 4 queries, 9 steps command and control, execution
  4. Agentic Identity and Cloud Control Plane Exploitation

    An intruder is using an agentic framework to automate cloud API interactions and identity access, characterized by successful sign-ins from unmanaged devices followed by high-tempo API bursts and local CLI tool usage.

    High Elastic Security Labs Part 1 of 2 · 3 queries, 8 steps command and control, execution
  5. ARTEX Automated Endpoint and Network Maneuvering

    An automated attacker is coordinating dual-use offensive tools like VShell, Chisel, or fscan to perform high-velocity internal discovery and credential harvesting.

    High Cisco Talos Part 2 of 2 · 3 queries, 8 steps command and control, credential access
  6. AI-Driven Identity and Cloud Exploitation

    An adversary is using an AI-orchestration framework to conduct high-velocity credential spraying against Entra ID and subsequently manipulate application credentials or directory roles for persistence at machine speed.

    High Cisco Talos Part 1 of 2 · 3 queries, 8 steps command and control, credential access
  7. Endpoint Exploit and Service Crash Monitoring

    An adversary is exploiting unpatched November 2026 vulnerabilities to gain access or elevate privileges, causing service crashes followed by anomalous child processes or suspicious RDP logins.

    High Cisco Talos 5 queries, 11 steps CVE-2026-105501, CVE-2026-105524
  8. Storm-3358: ARTEX AI Abuse and Cloud API Exploitation

    An adversary is using an agentic AI framework to perform rapid-fire credential spraying and API enumeration, leading to lateral movement via RDP and data exfiltration from compromised cloud identities.

    High Microsoft Part 2 of 2 · 5 queries, 11 steps command and control, credential access
  9. Suspicious Perimeter Reconnaissance and Web Exploitation

    An intruder is using automated agentic tools to identify exposed secrets on public-facing assets and is exploiting web vulnerabilities via command injection in user-agent headers to gain initial access.

    High Microsoft Part 2 of 2 · 3 queries, 8 steps defense evasion, discovery
  10. Agentic AI Cloud Resource Abuse (ARTEX)

    An adversary is using an agentic AI framework to automate cloud discovery and data exfiltration, identified by rapid, self-correcting API probing patterns and service principal abuse from campaign-associated infrastructure.

    High Microsoft Part 1 of 2 · 4 queries, 10 steps defense evasion, discovery
  11. LiteLLM Remote Code Loading and Exfiltration

    An adversary is using malicious Python startup hooks in a compromised package to execute a remote code loader and exfiltrate credentials from developer and production environments.

    High Microsoft Part 2 of 2 · 3 queries, 8 steps credential access, execution
  12. LiteLLM 1.107.0 Supply Chain Compromise and Credential Harvesting

    An intruder has distributed a compromised version of LiteLLM (1.107.0) that uses Python startup hooks to execute a credential harvester upon any Python invocation, targeting cloud, SSH, and browser secrets.

    High Microsoft Part 1 of 2 · 3 queries, 8 steps credential access, execution
  13. ConnectWise M365 Application Abuse

    An adversary has exploited compromised ConnectWise (SkyKick) vendor credentials to access M365 tenants, escalate directory privileges, and persist by adding secrets to secondary applications.

    Critical Huntress 4 queries, 10 steps credential access, initial access
  14. Ransomware Recovery Inhibition and Encryption

    An adversary is inhibiting system recovery by deleting volume shadow copies using standard tools or scripts before performing mass file encryption with rare extensions.

    Medium Huntress 4 queries, 9 steps impact
  15. Phishing Infrastructure and Interactive Exfiltration

    An adversary is operating the Iron Man System phishing kit to harvest credentials in real-time, using specific domains for infrastructure management and Telegram for automated exfiltration.

    High Microsoft Part 2 of 2 · 4 queries, 9 steps collection, command and control
  16. Cloud Tenant Hijacking and Service Principal Abuse

    An adversary uses stolen application secrets for application-only authentication to hijack M365 tenants, subsequently escalating privileges via the Graph API to host phishing infrastructure on SharePoint and OneDrive.

    High Microsoft Part 1 of 2 · 5 queries, 11 steps collection, command and control
  17. October 2026 Patch Tuesday Exposure Hunt

    Adversaries are exploiting newly disclosed October 2026 RCE vulnerabilities to trigger anomalous shell execution and deploy payloads on unpatched Windows and Exchange hosts.

    High Cisco Talos 3 queries, 7 steps CVE-2026-100202, CVE-2026-100209
  18. TeamFiltration Cloud Identity Spray and Pivot

    An adversary is using the TeamFiltration framework to spray M365 service accounts with default passwords from AWS infrastructure, subsequently harvesting data via the Graph API and probing internal VPN endpoints.

    High Proofpoint 5 queries, 11 steps collection, credential access
  19. TeamFiltration Service Account Compromise and Exfiltration

    An adversary has compromised functional service accounts via credential spraying to bypass MFA, then used these ghost identities to exfiltrate bulk data from M365 and discover administrative portals.

    High Proofpoint 4 queries, 10 steps credential access, discovery
  20. Storm-3068 Build Pipeline Execution and Tunneling

    An adversary has modified build pipelines to execute malicious code on agents, deploying RMM tools and establishing tunnels to exfiltrate Kubernetes credentials.

    High Microsoft Part 2 of 2 · 4 queries, 9 steps command and control, credential access
  21. Cloud Identity Takeover and DevOps Enumeration

    An adversary hijacked a cloud identity using self-service password reset to perform automated discovery across Azure DevOps repositories and harvest Kubernetes configuration files.

    High Microsoft Part 1 of 2 · 4 queries, 12 steps command and control, credential access
  22. Akira Ransomware Exfiltration and Impact

    An adversary is exfiltrating bulk data via SFTP using FileZilla and executing Akira ransomware, evidenced by massive outbound network transfers and the destruction of Volume Shadow Copies.

    Critical The DFIR Report Part 3 of 3 · 3 queries, 8 steps command and control, credential access
  23. Bumblebee Persistence and AD Credential Harvesting

    An adversary has established internal persistence through unauthorized remote access tools like RustDesk and is performing Active Directory credential harvesting by dumping the NTDS database and LSASS memory.

    High The DFIR Report Part 2 of 3 · 3 queries, 8 steps command and control, credential access
  24. Bumblebee Delivery and C2 Establishment

    An intruder has lured an administrator to a look-alike download page via SEO poisoning, leading to a trojanized installer that side-loads Bumblebee via consent.exe and establishes AdaptixC2.

    High The DFIR Report Part 1 of 3 · 4 queries, 11 steps command and control, credential access
  25. Microsoft Patch Tuesday September 2026 Exposure

    An adversary is exploiting September 2026 zero-day or critical remote code execution vulnerabilities, such as those in DNS Server or the Windows Update Stack, to establish initial access or escalate privileges on unpatched systems.

    High Cisco Talos 3 queries, 7 steps CVE-2026-58599, CVE-2026-65669
  26. Storm-3168 Web Application Probing

    An adversary is conducting automated web reconnaissance by probing for specific administrative and AI-related paths to identify vulnerable entry points for a subsequent cloud-focused intrusion.

    High Microsoft Part 2 of 2 · 3 queries, 8 steps credential access, discovery
  27. Storm-3168: Automated Azure Resource Destruction and Recovery Inhibition

    A compromised service principal is executing an automated sequence of Azure resource discovery, mass deletion, and credential collection to facilitate a cloud-native ransomware operation.

    High Microsoft Part 1 of 2 · 4 queries, 10 steps credential access, discovery
  28. Bumblebee Delivery and Persistence

    An intruder has delivered Bumblebee malware through a trojanized MSI installer via SEO poisoning, using DLL side-loading of consent.exe and establishing persistence with remote management tools like RustDesk.

    High The DFIR Report Part 1 of 2 · 4 queries, 11 steps credential access, execution
  29. Sideloaded AppX OAuth Token Theft

    An adversary has enabled Developer Mode and sideloaded a malicious AppX package to abuse WWAHost.exe, allowing them to capture MFA-compliant OAuth tokens via a legitimate Microsoft login dialog.

    High Huntress Part 1 of 2 · 3 queries, 8 steps credential access, defense evasion
  30. Bissa Scanner C2 and S3 Exfiltration

    An attacker is using Telegram for command-and-control alerts and Filebase S3 for data exfiltration after harvesting secrets from vulnerable application servers.

    High The DFIR Report Part 2 of 2 · 3 queries, 8 steps CVE-2025-55182, CVE-2025-9501
  31. Bissa Scanner Mass Exploitation and Credential Harvesting

    An attacker is using the Bissa scanner to exploit unauthenticated vulnerabilities in Next.js or WordPress, then harvesting sensitive credentials from .env files and cloud metadata.

    High The DFIR Report Part 1 of 2 · 3 queries, 10 steps CVE-2025-55182, CVE-2025-9501
  32. Rapid Phishing and Proxy-based Exfiltration

    An adversary has bypassed phishing-resistant MFA to gain initial access via a SaaS provider and is now using a multi-hop proxy or tunnel to exfiltrate data from an internal host.

    High Elastic Security Labs 3 queries, 10 steps command and control, exfiltration
  33. Rapid Identity Breakout and Exfiltration

    An adversary uses a compromised privileged identity to exfiltrate data via a multi-hop proxy or tunnel within 30 minutes of initial access, moving faster than traditional telemetry export batches.

    Medium Elastic Security Labs 3 queries, 10 steps command and control, exfiltration
  34. Cloud Workload Identity and Network Triage

    An adversary uses over-permissioned cloud identities to compromise container workloads, establishing persistence through rare outbound network channels that bypass standard scanning.

    Medium Microsoft Part 2 of 2 · 3 queries, 8 steps command and control, credential access
  35. Cloud Workload Runtime and Exploitation Behavior

    An adversary has exploited a public-facing containerized application and is maintaining persistence through binary drift or suspicious shell execution within the workload runtime.

    Medium Microsoft Part 1 of 2 · 3 queries, 11 steps command and control, credential access
  36. AD RMS Discovery and Administrative Reconnaissance

    An adversary is identifying on-premises AD RMS clusters via DNS and rights-policy template enumeration before escalating privileges via local group membership to reach the administrative surface.

    High Huntress 5 queries, 11 steps command and control, discovery
  37. Entra ID Agent User Impersonation and Teams Abuse

    An attacker uses the Entra ID Agent User OAuth flow to impersonate an AI agent and dispatch malicious content via Microsoft Teams using Graph API cmdlets.

    High Red Canary 3 queries, 8 steps execution, initial access
  38. Bulk Directory Discovery via AAD Graph API

    An adversary uses legacy Azure AD Graph API endpoints and known offensive Client IDs to perform bulk directory enumeration, specifically targeting internal API versions that expose sensitive authentication methods.

    High Elastic Security Labs 3 queries, 10 steps discovery, execution
  39. Shai-Hulud: Exfiltration and Deadman Switch

    An attacker is using the Shai-Hulud framework to exfiltrate stolen credentials through GitHub dead-drops and has installed a destructive deadman switch that triggers a home-directory wipe if tokens are revoked.

    High Datadog Security Labs Part 3 of 3 · 3 queries, 8 steps command and control, credential access
  40. Shai-Hulud Secret Harvesting and Discovery

    An intruder has deployed the Shai-Hulud framework to extract credentials from filesystem paths and process memory, subsequently using them to automate the discovery of cloud and Kubernetes infrastructure secrets.

    High Datadog Security Labs Part 2 of 3 · 3 queries, 8 steps command and control, credential access
  41. Shai-Hulud Framework Supply Chain Hook and Loader Bootstrap

    The adversary poisons a developer repository or AI coding assistant configuration to execute the Shai-Hulud loader and establish daemonized persistence.

    High Datadog Security Labs Part 1 of 3 · 3 queries, 10 steps command and control, credential access
  42. Suspicious ingress tool transfer via native utilities

    An adversary is using native Linux or macOS utilities like curl or wget to download malicious payloads from external infrastructure, hiding their activity within the high volume of legitimate cloud automation.

    Medium Elastic Security Labs 3 queries, 8 steps command and control
  43. AI Infrastructure Host Monetization and Persistence

    An attacker has compromised an AI gateway or retrieval engine and is now deploying masqueraded payloads to monetize the host via cryptomining and establish durable SSH or systemd persistence.

    High Microsoft Part 2 of 2 · 6 queries, 12 steps CVE-2025-68700, CVE-2026-24770
  44. AI Gateway Exploitation and Data Theft

    An intruder has exploited an exposed AI gateway or orchestration platform to harvest LLM API keys from process memory and exfiltrate tenant configurations from backend databases.

    High Microsoft Part 1 of 2 · 3 queries, 10 steps CVE-2025-68700, CVE-2026-24770
  45. CHAINDROP: C2 Discovery and Worm Propagation

    An attacker has infected local development environments via trojanized npm packages and is using Ethereum smart contracts to discover C2 infrastructure before propagating the worm using stolen GitHub credentials.

    High Elastic Security Labs Part 2 of 2 · 4 queries, 9 steps command and control, credential access
  46. CHAINDROP: Host-Based Node.js Worm Execution and Harvesting

    An attacker has gained initial access through a backdoored npm package preinstall hook, which executes a dropper to install a rogue Bun runtime and harvest developer credentials from local IDE configurations.

    High Elastic Security Labs Part 1 of 2 · 3 queries, 10 steps command and control, credential access
  47. Microsoft Graph and Cloud Application Exfiltration

    An adversary is using automated Graph API tools to enumerate organizational resources and exfiltrate SharePoint/OneDrive data after obtaining a cloud session via passkey-themed social engineering.

    High Microsoft Part 2 of 2 · 4 queries, 9 steps discovery, exfiltration
  48. TeamPCP Credential Validation and Discovery

    An adversary validates stolen cloud credentials and enumerates cloud infrastructure using offensive tools like TruffleHog or specialized Boto3 scripts following a supply chain compromise.

    High Wiz Part 1 of 2 · 4 queries, 9 steps discovery, execution

Hunt faster than the attacker.

Huntbase is the autonomous threat-hunting platform. When new research lands, Scout checks back through your history across the tools you already run, responds within the rules you set and verifies the threat is gone.

Content pack

Sign up to open

Pack contents open in Huntbase.