How gnilrats602/sentinelone-queries works
1. Repository purpose
- This is a content repository for SentinelOne Deep Visibility hunting queries, not an application or service.
- Its main role is to act as a structured catalog of detection logic that can be consumed by other tools for ATT&CK mapping, query navigation, and hunting workflows.
2. Primary artifact: YAML query documents
- The core unit is a .yml file describing one or more SentinelOne queries.
- Each file is metadata-rich: title, description, author, operating system, query text, MITRE tactic/technique/subtechnique, tags, false positives, and references.
- The schema is deliberately designed for programmatic parsing rather than manual-only reading.
3. Directory-level organization by platform
- Queries are grouped by operating system under queries/windows, queries/linux, and queries/apt.
- This creates the first architectural boundary: platform-specific detection content separated by target environment.
- Consumers can filter the repository by OS before further classification or execution.
4. Threat-model grouping
- The queries/apt folder holds campaigns and actor-oriented detections such as SolarWinds/SUNBURST-related content.
- This is a second content boundary: adversary/campaign-focused detections alongside OS-centric techniques.
- It complements the technique-centric Windows and Linux folders.
5. MITRE ATT&CK alignment as a cross-cutting layer
- The README makes clear that the repository is meant to support ATT&CK mapping.
- Tactic/technique/subtechnique fields provide a normalized taxonomy that external tooling can use to group, compare, and search detections.
- This is the main semantic layer connecting otherwise independent query files.
6. Query content and execution boundary
- The actual SentinelOne query syntax lives inside the query field of each YAML document.
- The repository does not include an execution engine, scheduler, or API integration; it is static source content.
- Any runtime interaction happens in downstream tools that read these files and submit queries to SentinelOne.
7. Template-driven consistency
- template.yml serves as the canonical shape for new query entries.
- This indicates a content-governance boundary: contributors should conform to a standard schema so downstream parsers remain stable.
- The template likely reduces drift across files and supports automation.
8. Data flow from authoring to consumption
- Content is authored or imported into YAML files, reviewed in the repo, and then consumed by other systems.
- Typical downstream uses mentioned in the README are:
- ATT&CK mapping
- a query navigator
- other hunting tools
- The repository therefore functions as a source-of-truth dataset for query intelligence.
9. Source lineage and curation state
- Many rules were converted from the older “SentinelOne ATTACK Queries” project.
- The README notes that some descriptions, references, and false positives still need cleanup, so the repo contains both curated and partially migrated content.
- This matters for diagramming because data quality is not uniform across all query files.
10. External dependencies that shape the architecture
- The content is inspired by SIGMA-style signatures and uses MITRE ATT&CK as the classification model.
- The detection language is SentinelOne Deep Visibility query syntax.
- These external standards define the structure of the files and the expectations of the consuming tooling.
11. What is not present
- There are no application services, databases, message queues, build pipelines, or deployment artifacts in the repository tree provided.
- So the architecture diagram should emphasize a documentation/content layer rather than a system runtime.
- The main components are repository structure, standardized YAML schema, and external consumers.
12. How to draw it
- A good diagram would show:
- the repo as a central query content store
- folders as platform/campaign partitions
- YAML files as individual detection records
- downstream tools as consumers for search, ATT&CK mapping, and query execution
- The most important boundary is between static curated detection content and external systems that operationalize it.
Main entry points: windows Windows · README.md README · linux Linux · apt APT
Components 19
Content store
Static detection content organized by platform and campaign.
- Repo · Source-of-truth repository for SentinelOne Deep Visibility hunting queries.
- README · Explains repository purpose, structure, and intended downstream uses.
README.md - Template · Canonical YAML shape for new query entries.
template.yml - Windows · Technique-focused Windows detection content.
queries/windows - Linux · Technique-focused Linux detection content.
queries/linux - APT · Adversary- and campaign-oriented detections.
queries/apt - PrintNightmare · Representative Windows hunting rule for a named technique or exploit pattern.
queries/windows/PrintNightmare.yml - Kerberoasting · Representative Windows credential-access hunting rule.
queries/windows/kerberoasting.yml - Sysmon Disable · Representative Windows defense-evasion hunting rule.
queries/windows/disable_sysmon.yml - Nix Sniffing · Representative Linux discovery or network-sniffing hunting rule.
queries/linux/nix_network_sniffing.yml - Firewall Off · Representative Linux defense-evasion or security-control tampering rule.
queries/linux/linux_firewall_disabled.yml - Sunburst · Campaign-focused content for SolarWinds/SUNBURST-related activity.
queries/apt/sunburst_campaign.yml
External standards
Schemas and taxonomies that shape the query documents.
- Deep Visibility · SentinelOne query syntax embedded in each YAML record.
- MITRE ATT&CK · Cross-cutting tactic/technique/subtechnique classification layer.
- SIGMA style · Signature-driven detection structure influencing the repository format.
Downstream consumers
Tools and workflows that read the repository content.
- ATT&CK map · Downstream mapping workflow that groups and compares detections.
- Query navigator · Downstream navigation and filtering workflow for hunting queries.
- Hunting tools · External tools that ingest YAML records for operational use.
- Query execution · Operational submission of parsed queries to SentinelOne by external systems.
Connections 26
- Repo → README: documents · Repository purpose and usage are explained here.
- Repo → Template: standardizes · Defines the canonical structure for new entries.
- Repo → Windows: partitions · Windows content is grouped separately.
- Repo → Linux: partitions · Linux content is grouped separately.
- Repo → APT: partitions · Campaign-oriented content is grouped separately.
- Windows → PrintNightmare: contains · Example query record within the Windows partition.
- Windows → Kerberoasting: contains · Example query record within the Windows partition.
- Windows → Sysmon Disable: contains · Example query record within the Windows partition.
- Linux → Nix Sniffing: contains · Example query record within the Linux partition.
- Linux → Firewall Off: contains · Example query record within the Linux partition.
- APT → Sunburst: contains · Example campaign-focused query record.
- Template → Deep Visibility: shapes · YAML records embed SentinelOne Deep Visibility query text.
- PrintNightmare → MITRE ATT&CK: annotated by · Queries carry ATT&CK tactic and technique metadata.
- Kerberoasting → MITRE ATT&CK: annotated by · Queries carry ATT&CK tactic and technique metadata.
- Nix Sniffing → MITRE ATT&CK: annotated by · Queries carry ATT&CK tactic and technique metadata.
- Sunburst → MITRE ATT&CK: annotated by · Campaign content is also ATT&CK-aligned.
- Repo → SIGMA style: inspired by · Repository structure follows signature-style detection conventions.
- README → ATT&CK map: enables · ATT&CK mapping is a stated downstream use.
- README → Query navigator: enables · Query navigation is a stated downstream use.
- README → Hunting tools: enables · Other hunting tools consume the catalog.
- Hunting tools → Query execution: submits · External tools operationalize queries against SentinelOne.
- ATT&CK map → MITRE ATT&CK: uses · Mapping workflow depends on ATT&CK taxonomy.
- Hunting tools → Deep Visibility: parses · Consumers read the embedded query syntax from YAML records.
- Query navigator → Windows: filters · Consumers can browse by platform partition.
- Query navigator → Linux: filters · Consumers can browse by platform partition.
- Query navigator → APT: filters · Consumers can browse by campaign partition.