Subscribe
Sign in
Home
Notes
Archive
Leaderboard
About
Prompt Injection Attacks: The Complete Threat Map for Production AI Systems
73% of production AI deployments are vulnerable. OpenAI calls it a ‘frontier, unsolved security problem.’ Here are all five attack vectors and the…
May 6
•
Owadokun Tosin Tobi
1
1
Most Popular
View all
What Are the 4 Biggest AI Agent Security Risks in 2026?
Mar 26
•
Owadokun Tosin Tobi
2
1
Anthropic's Battle Against AI Espionage:
Feb 23
•
Owadokun Tosin Tobi
1
2
Claude Hallucinated a GitHub Repo ID and Shipped It to Production.
Mar 5
•
Owadokun Tosin Tobi
1
4
LLM Hallucination in Production: Why Your Staging Environment Will Never Catch It
Mar 28
•
Owadokun Tosin Tobi
1
2
Recent posts
View all
AI Agent Security Framework: How to Build One Before Your First Production Deployment
Q1 2026 gave us four verified incidents, a new OWASP Top 10, and one lesson: the teams that got burned built agents before they built frameworks. Here…
Apr 13
•
Owadokun Tosin Tobi
1
2
LLM Hallucination in Production: Why Your Staging Environment Will Never Catch It
The model passed every evaluation. Then it hit production and invented a GitHub repo ID. Here is why that gap exists — and how to close it.
Mar 28
•
Owadokun Tosin Tobi
1
2
What Are the 4 Biggest AI Agent Security Risks in 2026?
88% of organisations reported a confirmed or suspected AI agent security incident last year. Here are the four risks driving those numbers with real…
Mar 26
•
Owadokun Tosin Tobi
2
1
AI Agent Security: The Complete Guide for Enterprise Builders in 2026
Every threat, every control, every real-world incident — from a production AI engineer who has built and broken these systems.
Mar 20
•
Owadokun Tosin Tobi
1
2
7 Fixes for the Indirect Prompt Injection Vulnerability That is Silently Leaking Agent Secrets
No Click Required. No Warning. Just Your Agent, a Link Preview, and an Attacker's Server.
Mar 17
•
Owadokun Tosin Tobi
1
1
Claude Hallucinated a GitHub Repo ID and Shipped It to Production.
These 5 Guardrails Would Have Stopped It.
Mar 5
•
Owadokun Tosin Tobi
1
4
When Agents Go Wrong:
How Anthropic, OpenAI, and Google Handle Agentic AI Edge Cases | A comparative breakdown of how the big three approach the hardest problems in…
Feb 27
•
Owadokun Tosin Tobi
1
1
1
Anthropic's Battle Against AI Espionage:
Inside the Detection of Industrial-Scale Distillation Attacks
Feb 23
•
Owadokun Tosin Tobi
1
2
See all
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
