Claude Code permission errors: what each message means
What each Claude Code blocking message means (bash denied by auto mode, deny rules, hooks) and the exact setting that unblocks it.
Library
All AI Coding Patterns
Technical articles: agents, architecture, tools and design decisions.
All articles
Prompt chaining vs chain of thought: inside or between calls
Chain of thought happens inside one model call. Prompt chaining, between several. Which to use, when to combine them, and the same example with both.
MCP vs Skills: differences and when to use each
MCP connects an agent to external tools in real time. A skill gives it packaged instructions it loads into its own behavior. When to use each one.
What is OpenSpec? The specs-before-code workflow
OpenSpec turns every change into a written proposal you approve before coding. What it is, how propose, apply and archive work, and when it pays off.
Claude Code's Permission Modes, Explained With Examples
Manual, acceptEdits, plan, auto, dontAsk and bypassPermissions: what each Claude Code permission mode does when it tries to delete a file or push to git.
Claude Code vs Antigravity CLI: Which One to Choose
Antigravity CLI replaced Gemini CLI in June 2026. I compare the two terminal agents against Claude Code and when each one makes sense.
MCP vs RAG: the differences and how they work together
MCP connects the model to live tools and data. RAG drops relevant text into the prompt. They do not compete: here is the difference and when to use both.
Agentic AI vs Generative AI: The Real Difference
Generative AI produces content from a prompt. Agentic AI adds a loop with memory and tools to complete multi-step tasks.
Cómo enseñar Git desde cero: un recorrido visual para el aula
Cómo enseñar Git desde cero en FP: secuencia de conceptos, ejemplos visuales y preguntas para comprobar qué entiende el alumnado.
Ejercicios de Git para principiantes: commits, ramas y equipo
Cuatro ejercicios de Git para principiantes con preparación, comandos y resultados esperados. Prácticas para alumnos de FP y aprendizaje individual.
Spec-driven development vs TDD and BDD: how they differ
TDD writes a test, BDD an example in business language and SDD a specification for the agent. What each one does and why you do not have to choose.
Claude vs ChatGPT for Coding: The Differences That Matter
Claude and ChatGPT don't solve the same problem when you code: one is built for autonomous agentic development, the other covers more ground outside of code. Which to use and when.