Skip to content
DocumentationAPI ReferenceChangelog
Talon ForgeTalon Studio v1.0.2-beta

Sigma & Native Queries

Author once in portable Sigma and let Talon translate it to every SIEM — or write native SPL, KQL, EQL or YARA-L directly.

Sigma — the default

Sigma is the canonical format: a portable, vendor-neutral rule you write once. The pySigma sidecar validates the YAML and compiles it to every connected target, so the same logic ships to each SIEM without a rewrite. Talon re-translates live as you edit and again when you save, storing the Sigma under queries.sigma and each translation beside it.

Sigma rule (queries.sigma)
title: Mimikatz LSASS Access
logsource:
product: windows
category: process_access
detection:
selection:
TargetImage|endswith: '\lsass.exe'
GrantedAccess: '0x1010'
condition: selection
level: high

Prefer a SIEM-specific rendering for one target? Click Customize on that target's preview — your text replaces the translation for that SIEM only, and Sigma stays the source for the rest. Reset to Sigma drops the override.

Native queries

When you'd rather write in your SIEM's own language — or need a feature Sigma can't express — author a native detection instead: Splunk SPL, Sentinel KQL, Elastic EQL or Chronicle YARA-L. It's stored verbatim in that dialect's field (queries.splunk, queries.sentinel, queries.elastic or queries.yara_l) and runs exactly as written.

Native SPL — Splunk only
index=windows EventCode=4688 New_Process_Name="*\\mimikatz.exe"
| stats count by ComputerName, Account_Name
Native rules don't cross-translate
A native query is not converted to the other SIEMs — it ships only to the platform whose language it's written in. Only Sigma is cross-compiled to all four targets.

Translation

The pySigma sidecar renders a Sigma rule per target, applying a processing pipeline where one exists so the output uses real field and source names rather than bare Sigma fields:

Splunk · SPL
pySigma's Splunk backend with the Windows pipeline (for example, service: security becomes source="WinEventLog:Security").
Sentinel · KQL
pySigma's Microsoft 365 Defender backend and pipeline, so the query targets the Defender advanced-hunting tables. Check it against the tables your Log Analytics workspace actually holds.
Elastic · EQL
pySigma's Elasticsearch EQL backend with the ECS Windows pipeline.
Chronicle · YARA-L
pySigma's Google SecOps backend, with no extra pipeline.
CrowdStrike · SentinelOne
Preview only: the Library shows these conversions, but Talon cannot deploy to them.

When a rule can't be expressed in a SIEM's language, that target shows an error instead of a query while the others still translate. Library rules are converted in advance and re-converted only when the rule itself changes.

Tip
Default to Sigma — it translates to all four SIEMs for free and keeps a detection portable across every environment you run. Reach for a native query only when you need a SIEM-specific capability Sigma can't express, and remember that rule then ships to that one platform alone.