Midnight Notes.
Collection of knowledge and ideas.
Concepts, experiments, and thoughts.
Highlighted notes
Testing strategy for AI-produced code
Testing strategies for AI-produced code: automation, Playwright, independent validation, regressions, and CI/CD.
Skills, MCP, RAG, and Memory: a practical example
Understand how Skills, MCP, RAG, and Memory complement one another in AI agents through the investigation of an HTTP 500 checkout error.
Context architecture, engineering, and documentation
Strategies to structure, organize, and maintain project technical context for developers and AI agents.
Explore by topic
Context, memory, and tools
How to organize knowledge and connect the tools that help agents work.
Planning and delegation
How to turn goals into clear tasks and distribute responsibilities between humans and agents.
Parallelism and orchestration
How to coordinate agents, manage dependencies, and integrate work performed in parallel.
Code and quality
How to investigate projects, evolve implementations, and evaluate AI-produced code.
Security and access
How to protect information, control permissions, and assess application security.
Delivery and operations
How to take changes to production, monitor results, and recover operations when something fails.
Last notes
Development and delivery workflow
How to integrate AI agents into CI/CD to prepare pull requests, investigate failures, and support releases and deployments.
Code review in the AI era
Strategies for code review in the AI era: independent reviewer agents, evidence, and risk-oriented human review.
Testing strategy for AI-produced code
Testing strategies for AI-produced code: automation, Playwright, independent validation, regressions, and CI/CD.