Bot Radar & Risk Dashboard
Relevant source files
The following files were used as context for generating this wiki page:
The Bot Radar and Risk Dashboard systems within FusionOps serve as the primary defensive layer for affiliate landing pages and advertising accounts. Bot Radar provides real-time visibility into bot traffic (specifically Google Ads and Search bots), utilizing a 4-signal detection mechanism to identify anomalies before they result in account bans. The Risk Dashboard (driven by the FBIS Hermes system) aggregates intelligence from multiple autonomous agents to calculate risk scores and execute auto-pause rules for accounts deemed "Critical."
Sources: src/components/BotRadar.jsx:291-294, HERMES_FBIS_AUDIT.txt:104-118
Bot Radar System
The Bot Radar is a real-time monitoring component that tracks bot visits across unique sites over various time intervals (1h, 24h, and 7d). It identifies bot types such as adsbot, googlebot, inspection, and verification using a combination of IP CIDR matching, Reverse DNS (PTR), User-Agent analysis, and Cloudflare Bot Scores.
Sources: src/components/BotRadar.jsx:45-55, src/components/BotRadar.jsx:215-225
Detection Signals and Weights
The system evaluates traffic using four distinct signals, each contributing to an overall anomaly score:
| Signal | Description | Weight |
|---|---|---|
| IP CIDR Match | Matches the requester IP against known bot IP ranges. | 35% |
| Reverse DNS | Performs PTR match to verify the hostname belongs to a valid bot provider. | 35% |
| User-Agent | Validates the User-Agent string for bot identifiers. | 20% |
| CF Bot Score | Utilizes Cloudflare's native behavioral bot scoring. | 10% |
Sources: src/components/BotRadar.jsx:218-223
Pre-Ban Anomaly Scoring
The Bot Radar identifies "Pre-Ban Signals" by calculating an anomaly_score for site domains. If the anomaly score exceeds 0.7, the site status is elevated to "Critical," indicating a high probability of an impending account ban.
The diagram shows the flow from traffic detection in the Pixel Worker to the generation of pre-ban alerts based on weighted detection signals.
Sources: src/components/BotRadar.jsx:167-175, CLAUDE.md:65-70
FBIS Risk Scoring System
The FusionOps Ban Intelligence System (FBIS) is an autonomous agent-based framework that performs nightly risk analysis on all active accounts. It utilizes the FBIS MCP Server (Python FastMCP) to interact with API Workers and D1 databases.
Sources: HERMES_FBIS_AUDIT.txt:13-25, CLAUDE.md:82-87
Risk Scoring Agents
The system orchestrates five specialized Hermes agents to generate a comprehensive risk verdict:
- ARGUS: Analyzes Proxy and IP risk.
- NEXUS: Evaluates account isolation scoring to prevent cross-contamination.
- IRIS: Monitors traffic quality through pixel event analysis.
- CHRONO: Predicts ban timelines based on historical patterns.
- VERDICT: Acts as the aggregator, calculating a weighted average from the other agents.
Sources: HERMES_FBIS_AUDIT.txt:46-55, HERMES_FBIS_AUDIT.txt:104-110
Scoring Scales and Actions
Risk scores are mapped to a 0-100 scale, where higher scores indicate higher risk.
| Score Range | Status | Action Required |
|---|---|---|
| 0-30 | 🟢 HEALTHY | No action; safe for operation. |
| 31-60 | 🟡 WATCH | Monitor closely for anomaly shifts. |
| 61-80 | đź”´ RISK | Action required; reduce budget by 50%. |
| 81-100 | đź”´ CRITICAL | Immediate account pause via Auto-Pause Rules. |
Sources: HERMES_FBIS_AUDIT.txt:111-125
Analysis Pipeline Architecture
The risk analysis follows a nightly pipeline triggered via the /fbis-run skill.
The sequence diagram illustrates the transition from parallel agent analysis to serial verdict aggregation and subsequent alerting.
Sources: HERMES_FBIS_AUDIT.txt:68-85, hermes-agents/INDEX.md:12-25
Integration and Infrastructure
The Bot Radar and Risk Dashboard rely on a multi-database Cloudflare D1 setup and the FusionOps API Worker.
Database Requirements
- fusionops-pixel-new-v2: Stores tracking events, clicks, and conversions used by IRIS and Bot Radar.
- fusionops-main-new-v2: Contains campaign and account metadata.
- risk_scores: A specific D1 table created for persistent storage of agent findings.
Sources: CLAUDE.md:100-105, INFRASTRUCTURE_SETUP.md:144-155, HERMES_FBIS_AUDIT.txt:161-167
API Integration Points
The systems interact via the following endpoints:
GET /api/bot-radar/summary: Retrieves aggregated bot visit totals.GET /api/bot-radar/pre-ban: Returns sites with elevated anomaly scores.POST /api/risk: Used by the FBIS VERDICT agent to write calculated scores.GET /api/pixel-events: Queried by IRIS for traffic quality analysis.
Sources: src/components/BotRadar.jsx:258-265, HERMES_FBIS_AUDIT.txt:152-159
Summary
The combination of the Bot Radar's real-time behavioral detection and the Risk Dashboard's nightly autonomous agent analysis provides FusionOps with a proactive defense mechanism. By automating the identification of bot traffic patterns and cross-referencing them with account isolation and proxy integrity, the system significantly reduces the operational impact of advertising platform bans.
Sources: HERMES_FBIS_AUDIT.txt:190-205, src/components/BotRadar.jsx:290-295