AI agent security for builders who ship agents — not slideware. Prompt injection, MCP security, AI coding agent attacks, poisoned READMEs, malicious hooks, and tool poisoning: broken down with mock UIs, then fixed with concrete controls.
What you get:
• Short attack teardowns (how it works → what breaks → the fix)
• Long-form deep dives with a clear "My call" verdict from a security engineer
• Defence checklists mapped to real agent/MCP failure modes
• Fast incident reactions when a new agent CVE or campaign drops
Cadence: Shorts most weekdays · 1–2 long-form deep dives per week.
Scripts by Jayesh K S (India) — agent-security engineer. Narration is AI-generated from those scripts. Visuals are simulations; payloads are redacted or non-functional. Opinions in "My call" are mine — not vendor endorsements.
#AIsecurity #PromptInjection #AIagents #MCP #AppSec
Joined 8 October 2026