Open-source AI Security Operations Center: alert fusion, LLM-agent triage, MITRE ATT&CK investigation, and a replayable decision ledger for every agent step. Self-hostable, runs with no API keys, MIT licensed. Ships an MCP server for Claude, Cursor and Continue.
Python agentic-aiai-securityalert-triageblue-teamcybersecuritydetection-engineeringincident-responselanggraphllm-agentsmcpmitre-attackmodel-context-protocolpurple-teamsecurity-operationsself-hostedsiemsoarsocthreat-detectionthreat-intelligence
"""Natural language → query translation for AiSOC.
This module ships a deterministic, offline-friendly translator that converts
plain-English security questions into ES|QL (OpenSearch / Elasticsearch),
KQL (Microsoft Sentinel / Kusto-style), and SPL (Splunk) dialects. It also
exposes a grammar validator so consumers can refuse to execute syntactically
invalid output.
Design notes:
- The translator runs **without** an LLM by default. This is what allows the
air-gapped deployment story to keep working and lets the eval harness give
reproducible numbers in CI.
- An optional LLM enhancement path is exposed via :func:`enhance_with_llm` for
callers that already have credentials available; if it fails for any reason
we fall back to the deterministic translation.
- Every emitted query goes through grammar validation before being returned,
and the API layer also re-validates before execution.
"""
from .grammar import (
GrammarError,
validate_esql,
validate_kql,
validate_spl,
)
from .translator import (
NLQuery,
QueryIntents,
TranslatedQuery,
Translator,
enhance_with_llm,
parse_intents,
translate,
)
__all__ = [
"NLQuery",
"QueryIntents",
"TranslatedQuery",
"Translator",
"translate",
"parse_intents",
"enhance_with_llm",
"GrammarError",
"validate_esql",
"validate_kql",
"validate_spl",
]