BAND HACKATHON 2026 — TRACK 3: REGULATED WORKFLOWS

AI Agents That

Investigate Cyber Threats

|

Phishing Email

Suspicious Login

File Exfiltration

Malware Alert

HOW IT WORKS

Six specialized AI agents collaborate in real time through Band — each posting findings to a shared investigation room.

01INTAKE AGENT

Threat Detected

Submit a suspicious email, file, or message. The Intake Agent instantly classifies the threat type and severity.

02THREAT + EVIDENCE AGENTS

Deep Analysis

Threat Intelligence Agent scans domains and indicators. Evidence Agent cross-references login history and builds a timeline.

03RISK + RED TEAM AGENTS

Agents Debate

Risk Agent calculates severity. Red Team Agent challenges the weakest evidence. Risk Agent defends or revises — live debate visible.

04RESPONSE AGENT

Action Plan

A prioritized containment plan and executive summary is generated — ready for human review and approval.

05HUMAN REVIEWER

You Decide

Critical actions require your approval. Every decision is logged in a full audit trail for compliance and traceability.

BUILT WITH

Kestrion runs on Band for multi-agent coordination, an AI/ML reasoning layer, and a Next.js command center.

AGENT ORCHESTRATION

Band SDK

Coordinates all six agents and manages the shared investigation room where findings are posted in real time.

ROOM + APPROVALS

Band REST API

Creates incident rooms, routes agent messages, and handles human approval and audit-trail logging.

REASONING ENGINE

AI / ML API

Powers the language reasoning behind every agent — classification, debate, risk scoring, and response planning.

FRONTEND

Next.js

The React command-center interface, live agent feed, and this landing experience.

6

AI Agents

99%

Threat Confidence

<5min

Investigation Time

100%

Human Approval

READY TO INVESTIGATE?

Submit any suspicious email, login alert, or security incident and watch agents investigate in real time.

BAND SDK · BAND REST API · AI/ML API · NEXT.JS

BUILT BY TEAM SHAHEEN FOR BAND HACKATHON 2026