Threat Intelligence for the Application Layer
Public-source threat intelligence on the API, AI and business-logic attack surface. Built for security teams who'd rather know now than read about it later.
LATEST REPORTS
The 2025–2026 threat landscape, three angles.
APIsec Labs
AI Surface · September 2026
State of AI Surface
AI Attack Surface Research Report · Expanded Edition
50
Repos Scanned
7,615
AI Surfaces
95%
Runtime-Only
96%
Trip a Gov Rule
50%
Expose MCP
The expanded edition: the open-source ai-surface scanner run across 50 of the most popular AI projects on GitHub, about 2.6M combined stars, cataloging 7,615 AI attack surfaces. The biggest AI attack surface is not the model, it is the ordinary REST API underneath it. 5,481 of those surfaces are conventional HTTP endpoints and 95% of everything found can only be confirmed safe or exploitable against a running application. An exposure report, not an incident report.
Get the reportAPIsec Labs
AI Surface · June 2026
State of AI Surface
AI Attack Surface Research Report
19
Repos Scanned
100%
Trip a Gov Rule
83%
Ship AI Agents
67%
BOLA-Candidate APIs
42%
Expose MCP
The open-source ai-surface scanner run across 19 of the most popular AI projects on GitHub, about 941k combined stars. Agents and RAG are now the baseline, the biggest AI attack surface is still ordinary BOLA-prone REST, and 100% of the applications would trip an AI-governance rule. An exposure report, not an incident report: what is being shipped into the application layer, before anyone attacks it.
Get the reportAPIsec Labs
API Security · 2025–2026
API-Layer Attacks
API Security Threat Intelligence Report
14
Critical CVEs
59%
Zero-Auth Vectors
97%
Exploitable ≤1 Request
400%
AI/API Overlap Growth
30
Incidents
Thirty high-impact API incidents dissected — auth bypasses, BOLA, unauthenticated RCE and AI-supply-chain compromise. The briefing your security team actually needs before the next one lands.
Get the reportAPIsec Labs
AI Security · 2025–2026
AI-Focused Attacks
AI Security Threat Intelligence Report
9
Critical CVEs
210%
AI Vuln Growth
540%
Prompt Inj Growth
$100K+
Max LLMjacking/Day
15
Incidents
Fifteen AI-focused incidents covering prompt injection, model theft, training data poisoning, and AI-supply-chain attacks. The pace of vulnerability disclosure is accelerating faster than defensive tooling.
Get the reportAPIsec Labs
Business Logic · 2025–2026
Business Logic Exploits
Business Logic Security Intelligence Report
100%
Scanner-Invisible
59%
BL Attack Growth
27%
Of All API Attacks
45%
Crypto Bounty Share
15
Incidents
Fifteen documented business-logic exploits across e-commerce, financial services, and crypto. 100% of these attacks are invisible to traditional scanners — the flaw lives in how the application is designed to work, not in how it's coded.
Get the reportKEEP THE PULSE
New reports drop every two months.
Get future Industry Pulse reports the day they're published, alongside research, advisories and tool updates from APIsec Labs.
All incidents sourced from public CVE databases, vendor advisories, responsible-disclosure publications, and peer-reviewed security research.