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
import clsx from 'clsx';
import type { EvalReport } from './BenchmarkResults';
/**
* 2026 published KPI bar — the buyer-side scoring rubric AiSOC commits to.
* Mirrors `DEFAULT_KPI_BAR_TARGETS` in `services/api/app/services/sla.py`
* and the four metrics surfaced in the in-app SLA dashboard
* (`apps/web/src/components/sla/SLADashboard.tsx`).
*
* MTTD/MTTR are tenant-runtime metrics, surfaced inside the product rather
* than on the public scoreboard. The four metrics below are the ones we can
* report from the open synthetic harness without needing a live tenant.
*/
interface KpiTarget {
id: string;
label: string;
target: string;
/**
* Live observed number from the eval report when available, else "n/a".
*/
observed?: string;
observedPasses?: boolean;
blurb: string;
}
function fmtPct(value: number): string {
return `${(value * 100).toFixed(1)}%`;
}
function buildTargets(report: EvalReport): KpiTarget[] {
const alertReduction = report.suites.alert_reduction;
const mitre = report.suites.mitre_accuracy;
// Alert-to-incident ratio is derived from the alert-reduction suite:
// 1,000 fabricated alerts compress to N incidents → ratio = 1000 / N.
const alertsIn = (alertReduction?.details?.alerts_in as number | undefined) ?? 1000;
const incidentsOut = alertReduction?.details?.incidents_out as number | undefined;
const ratio = incidentsOut && incidentsOut > 0 ? alertsIn / incidentsOut : undefined;
const ratioObserved = ratio ? `${ratio.toFixed(1)}:1` : undefined;
const ratioPasses = typeof ratio === 'number' ? ratio >= 50 : undefined;
// MITRE technique tagging: per-template macro is the closest published
// proxy on the open harness. The in-tenant SLA dashboard uses the live
// `mitre_technique_tagging_min_pct` from `tenant_sla_summary`.
const macro = mitre?.per_template;
const macroValue = macro?.value;
const macroPasses = typeof macroValue === 'number' ? macroValue >= 0.85 : undefined;
return [
{
id: 'fp_rate',
label: 'False-positive rate',
target: '≤ 5%',
blurb:
"A live tenant metric from `services/api`'s SLA service — surfaced in the in-app SLA dashboard, not on the public harness. The harness can't measure FP rate against synthetic data.",
},
{
id: 'a2i_ratio',
label: 'Alert-to-incident ratio',
target: '≥ 50:1',
observed: ratioObserved,
observedPasses: ratioPasses,
blurb:
'Derived from the alert-reduction suite: a 1,000-alert noisy stream is fed through the in-harness fusion grouping, and the ratio is 1,000 ÷ incidents_out.',
},
{
id: 'mitre_t',
label: 'MITRE technique tagging',
target: '≥ 85%',
observed: typeof macroValue === 'number' ? fmtPct(macroValue) : undefined,
observedPasses: macroPasses,
blurb:
'Per-template macro accuracy across 55 incident templates is the closest open-harness proxy. In-tenant data uses the live ECS technique-tag rate.',
},
{
id: 'mitre_st',
label: 'MITRE sub-technique tagging',
target: '≥ 60%',
blurb:
'Live tenant metric. The synthetic harness labels at the tactic level; sub-technique coverage is reported from live detection content in the in-app SLA dashboard.',
},
];
}
interface KpiBarProps {
report: EvalReport;
}
export function KpiBar({ report }: KpiBarProps) {
const targets = buildTargets(report);
return (
<div className="rounded-2xl border border-white/10 bg-white/[0.02] p-6">
<div className="flex flex-wrap items-baseline justify-between gap-3">
<div>
<h3 className="text-lg font-semibold tracking-tight text-white">
2026 published KPI bar
</h3>
<p className="mt-1 max-w-2xl text-sm text-gray-400">
The four buyer-side targets AiSOC commits to. The two derivable
from the open harness are checked live below; the other two are
tenant-runtime metrics surfaced in the in-app SLA dashboard.
</p>
</div>
<a
href="https://github.com/beenuar/AiSOC/blob/main/services/api/app/services/sla.py"
target="_blank"
rel="noreferrer"
className="text-xs text-gray-500 underline decoration-dotted hover:text-gray-300"
>
Source: services/api/app/services/sla.py
</a>
</div>
<dl className="mt-5 grid gap-3 sm:grid-cols-2 lg:grid-cols-4">
{targets.map((t) => {
const hasLive = typeof t.observed === 'string';
const passes = t.observedPasses;
return (
<div
key={t.id}
className="rounded-lg border border-white/5 bg-black/20 p-4"
>
<dt className="text-[11px] font-semibold uppercase tracking-wider text-gray-500">
{t.label}
</dt>
<dd className="mt-1 flex items-baseline gap-2">
<span className="font-mono text-xl font-semibold tabular-nums text-white">
{t.target}
</span>
{hasLive && (
<span
className={clsx(
'inline-flex items-center gap-1 rounded-full border px-2 py-0.5 text-[10px] font-semibold',
passes === false
? 'border-rose-500/30 bg-rose-500/10 text-rose-300'
: passes
? 'border-emerald-500/30 bg-emerald-500/10 text-emerald-300'
: 'border-white/10 bg-white/[0.03] text-gray-400',
)}
>
<span
className={clsx(
'h-1.5 w-1.5 rounded-full',
passes === false
? 'bg-rose-400'
: passes
? 'bg-emerald-400'
: 'bg-gray-400',
)}
/>
Live: {t.observed}
</span>
)}
{!hasLive && (
<span className="text-[10px] uppercase tracking-wider text-gray-600">
in-app
</span>
)}
</dd>
<p className="mt-2 text-xs leading-relaxed text-gray-500">
{t.blurb}
</p>
</div>
);
})}
</dl>
</div>
);
}