AI Risk Scanner

An OpenWatch Project

Sample Report — Based on a real scan, redacted to protect the source

Deep AI Security Risk Analysis

Executive Summary

xxxxxxxx was scanned on xxx x, 2026 at 00:07:13 UTC. The assessment identified 19 total findings across multiple security domains: 12 informational, 4 medium-severity, 2 high-severity, 1 low-severity, and 0 critical findings. While the organization demonstrates solid baseline security practices including proper email authentication and HTTPS enforcement, several concerning exposures create attack pathways that require immediate attention.

The most significant concerns are the presence of protected source control files xxxxxx and xxxxxxx in the web root, an exposed employee portal accessible without authentication, and the disclosure of internal infrastructure paths through robots.txt. These findings, combined with weak Content Security Policy directives, create a concerning attack surface that could enable credential compromise and lateral movement.

Risk Score

HIGH — Two high-severity findings involving potential source control and secrets exposure, plus multiple medium-severity configuration weaknesses.

Email Security

Example Company maintains strong email security controls:

Overall email security posture is strong with properly configured anti-spoofing controls.

Domain & Certificate Exposure

Certificate transparency logs reveal 23 unique subdomains with 11 currently live. Notable findings include:

Sensitive exposure: staging.xxxxxxxxx.xxx is publicly visible in CT logs
Email infrastructure: 4 email service subdomains detected (email.xxxxxxx, email.xxxxx, email.xxxxxxx, email.support)
Active portals: portal.xxxxxxxxx.xxx and app.xxx.xxxxxxxxxx are accessible

The presence of staging.xxxxx.xxxxx. in public certificates represents information disclosure that could aid reconnaissance efforts.

Code & Secret Exposure

GitHub scanning found no public repositories containing references to xxxxxxxxxxxxxxxx. No AI framework code, hardcoded secrets, or API keys were discovered in public code repositories. The organization maintains clean public code hygiene.

Breach History

No historical data breaches involving xxxxxxxxxxxxxxx were found in Have I Been Pwned records. This indicates no known credential leakage from third-party breaches affecting the primary domain.

Web Attack Surface

Web discovery identified several concerning exposures:

Path Status Code Severity Description
/robots.txt 200 Medium Reveals sensitive paths (/admin/, /private/, /api/, /employee/)
/.git/HEAD 403 High Source control directory present but protected
/.env 403 High Environment file present but protected
/employee 200 Medium Employee panel accessible without authentication
/sitemap.xml 200 Info Contains 24 URLs

Critical Findings:

AI-Related Findings: No embedded AI chatbots or client-side AI API calls were detected.

HTTP Security Posture

Header Status Value / Issue
HSTS Present max-age=31536000; includeSubDomains; preload
CSP Weak Contains unsafe-inline and unsafe-eval
X-Frame-Options Present SAMEORIGIN
X-Content-Type-Options Present nosniff
Referrer-Policy Present strict-origin-when-cross-origin
Permissions-Policy Present camera=(), microphone=(), geolocation=()

Security Header Score: 6/6 — All major security headers are present.

Key Issues:

Positive Controls:

Known Vulnerabilities (CVE)

No CVEs were identified. Technology fingerprinting detected only Cloudflare as the primary identifiable service, with no vulnerable product versions exposed.

Threat Intelligence

xxxxxxxxxxxxxxxxxxx shows no associations with known threat actors or malicious campaigns in OTX threat intelligence. All resolved IP addresses (104.xx.xx.xxx, 104.xx.xx.xxx, and IPv6 equivalents) are clean with zero threat pulse matches.

Network Exposure

Shodan identified 12 open ports on 104.xx.xx.xxx (80, 443, 2053, 2082, 2083, 2086, 2087, 2095, 2096, 8080, 8443, 8880). All services appear to be Cloudflare-managed, significantly limiting direct exploitation risk. The host is properly protected behind Cloudflare’s CDN infrastructure.

No exposed databases, AI/ML services, or unprotected administrative interfaces were found on the public IP.

AI Attack Surface

No significant AI-specific attack surface was identified:

The organization appears to have minimal public-facing AI infrastructure, reducing AI-specific attack vectors.

Attack Scenario

An attacker would begin by leveraging the information disclosure in robots.txt, which reveals the existence of /admin/, /private/, /api/, and /employee/ endpoints. The attacker would then target the accessible /employee portal (HTTP 200), which requires no authentication and likely contains employee-facing functionality or login interfaces.

Using the robots.txt intelligence about /api/ endpoints, the attacker would attempt to enumerate API paths and potentially discover administrative functions. If the employee portal contains login functionality, the attacker would attempt credential stuffing attacks using common business automation industry passwords, given the company’s focus on xxxxxxx xxxxxxxxx services.

The protected but present .git and .env files (returning 403) suggest these sensitive files exist in the web root — if access controls fail or misconfigurations occur during updates, these would immediately expose source code and environment variables including database credentials, API keys, and other secrets. The weak CSP with unsafe-inline and unsafe-eval directives would facilitate any XSS attacks needed to steal administrator credentials from the employee portal.

This scenario would ultimately provide administrative access to business automation systems and customer data, enabling data theft, system manipulation, or ransomware deployment across client integrations.

AI Adoption & Supply Chain

Job posting analysis could not be completed due to insufficient data retrieval for “xxxxxxxxxxxxxxx” searches. Manual review of the company’s LinkedIn careers page and website content suggests focus on business automation and process optimization, but specific AI technology stack details remain unavailable.

The website content mentions “AI policy” in the sitemap, suggesting some level of AI governance awareness, but AI adoption patterns cannot be assessed without job posting data.

Recommendations

Immediate (24–72 hours)

Short-term (30 days)

Strategic (90 days)

Methodology

This assessment analyzed xxxxxxxxxxxxxxx using DNS enumeration, GitHub scanning, breach database queries, Shodan network reconnaissance, web discovery probing, HTTP security analysis, and threat intelligence correlation. Data was collected on xxxxx xx, 2026 at 00:07:13 UTC.

Known Limitations: Certificate transparency collection failed, preventing subdomain enumeration. Job posting analysis was unsuccessful, limiting shadow AI assessment. This assessment covers public-facing attack surface only and does not include internal network security, application logic flaws, or social engineering vectors.

Coverage: Email security, web application security, network exposure, code repositories, breach history, and basic AI attack surface analysis for public infrastructure only.