Andy Agent Lab

Articles

  1. Beyond the prompt: AI models, reasoning, and speed

  2. How a Forged Message Can Hijack an AI Workflow

  3. Cognitive Closure in AI Use: Why We May Stop Too Soon

  4. How to Write for AI Models: The Language of Effective Prompts

  5. Role-Play Jailbreaks: A Testable False-Premise Mechanism

  6. Can AI Lie? The Difference Between Error, Hallucination, and Deception

  7. LLM vs. AI System Capabilities: What Architecture Adds

  8. Human Capabilities vs. AI Behavior: 26 Comparisons

  9. LLM Improper Output Handling: Risks, Controls, and Test Cases

  10. When AI Systems Optimize Against the User’s Real Goal

  11. User Prompt vs. Model Context: What an LLM Actually Receives

  12. LLM vs. RAG vs. AI Agents vs. MCP: Architecture Roles

  13. LLM Agent Memory Architecture: Types, Lifecycle, and Evaluation

  14. LLM Prompt Injection: OWASP Boundary Failures and Controls

  15. AI Agent Security Audit: 8 Trust-Boundary Checkpoints

  16. How AI Tools Read Emotional Signals in Text

  17. Gmail and WhatsApp AI Agents: Private-Message Security Risks

  18. AI File Upload vs. Full-File Review: How to Verify Coverage

  19. Why Clowns and AI-Generated Content Can Feel Uncanny

  20. Vibe Coding Risks: Code Defects and Verification Gaps

  21. Observed Classification Layers in ChatGPT

  22. Parallel Reasoning in LLM Systems: Orchestration Pattern

  23. Connected Apps and MCP Security: Permissions, Data, and Actions

  24. Web Retrieval Prompt Injection Boundary in LLM Systems

  25. Theory of Mind in LLMs: How Models Track Beliefs, Intentions, and Perspectives

  26. LLM Sycophancy: Definition, Evidence, and Evaluation

  27. Prompt Engineering for Reliable AI Work

  28. Orders of Intentionality in LLM Evaluation

  29. Tool-Using LLMs: Model-Led vs. Orchestrator-Led Execution

  30. LLM Memory vs. Context: 5 Architecture Boundaries

  31. Request Assembly Threat Model for AI Agents

  32. LLM Boundary Assurance Failures: Client-Captured Security Report

  33. LLM Prompt Assembly Security: Policy and Untrusted Data

  34. LLM Integration Trust Boundaries: Threat Modeling Before AI Agents

  35. AI Agent Orchestration Loops: Security Risks and Controls

  36. Social Engineering in AI Systems and Decision Pipelines

  37. Tool-Using LLM Systems: Privilege Bleed and Integrity-Signal Failures

  38. Human vs. GenAI Capabilities: 22 LLM Gaps and Engineering Controls

  39. LLM Fluency vs Factuality: Why Fluent Answers Can Be Wrong

Get in touch

info@andyagentlab.com