Introduction to LLM
This page provides an easy-to-understand guide on LLMs (Large Language Models) from basics to applications for AI enthusiasts.
Chapter 11 — Attack Surfaces and Protocol Vulnerabilities
Eleventh post of the LLM Primer IV walkthrough. The classical attacks adapted to MCP — Confused Deputy, Token Passthrough, Session Hijacking — the protocol-level flaws around capability escalation and unauthenticated sampling, and the implicit trust propagation that makes context poisoning a structural problem rather than a hygiene one.
2026-04-09Chapter 10 — Long-Horizon Task Memory
Tenth post of the LLM Primer IV walkthrough. Short-term memory through windows and ReAct scratchpads, long-term memory through episodic vectors and semantic stores, and the compaction techniques that keep an agent productive over hours and days.
2026-04-08Chapter 8 — Architectural Deployment Layouts
Eighth post of the LLM Primer IV walkthrough. The three deployment layouts that have emerged in the MCP ecosystem — reusable agent, strict purity, hybrid — and the four binding constraints that determine which one fits which project.
2026-04-06Chapter 7 — Advanced Collaborative and Dynamic Patterns
Seventh post of the LLM Primer IV walkthrough. Roundtable consensus, handoff routing, and magentic orchestration — the patterns that emerge when the topology has to be built per request, with the failure modes (non-termination, mis-routing, runaway planning) the simpler patterns avoid.
2026-04-05Chapter 5 — Transport Protocols and Discovery
Fifth post of the LLM Primer IV walkthrough. The three transports MCP supports, the .well-known discovery layer with Server Cards, and the boring operational concerns — CORS, origin validation, caching — that decide whether a server is a cooperative network citizen or a liability.
2026-04-03Chapter 4 — Client Primitives: Agentic Behaviors and Control
Fourth post of the LLM Primer IV walkthrough. Sampling, Roots, and Elicitation are the three small, controlled holes MCP punches through the host-server wall — each a capability granted back, each a risk accepted on the user's behalf.
2026-04-02
Chapter 3 — Server Primitives: Exposing Context and Capabilities
Third post of the LLM Primer IV walkthrough. The three nouns an MCP server can offer — Resources (read state), Prompts (reusable scaffolding), Tools (write actions) — their schemas, their lifecycles, their error models, and the discipline of choosing the right primitive.
2026-04-01
Chapter 2 — Unveiling the Model Context Protocol (MCP)
Second post of the LLM Primer IV walkthrough. What MCP actually standardizes, the three-role split of Host, Client, and Server, why dynamic discovery and bidirectional messaging differ from REST in the cases that matter, and the session lifecycle that opens with capability negotiation.
2026-03-31
Chapter 1 — The AI Integration Crisis and the Rise of Agentic Architecture
First post of the LLM Primer IV walkthrough. Why monolithic agents fray as system prompts grow, the N times M integration problem hiding underneath, and the move from prompt engineering to context engineering that MCP was built to enable.
2026-03-30
LLM Primer IV — Series Introduction & Index
Kicking off the chapter-by-chapter walkthrough of Book IV in the LLM Primer series — Designing AI Cognition with MCP. Why agents need a protocol layer to scale past demoware, who this book is for, and the schedule for the fourteen posts that follow, March 30 through April 12.
2026-03-29
Chapter 6 — RAG Threat Models and Vulnerabilities
Sixth post of the LLM Primer III walkthrough. The expanded attack surface of retrieval — corpus poisoning, adversarial chunks, indirect prompt injection, embedding inversion, and the confused-deputy problem in agentic RAG. Concrete attacks, each demonstrated, each reproducible.
2026-03-23