My Claude Code Skills in Production: What the Audit Taught Me
9 skills in production. The audit put them under the microscope: length, salience dilution, long-context drift. What holds, what falls apart, and why the search engine changes discovery.
Lessons learned, problems and solutions
9 skills in production. The audit put them under the microscope: length, salience dilution, long-context drift. What holds, what falls apart, and why the search engine changes discovery.
After the JavaScript basics, the next step up: classes, the this trap and prototypes, modules, Web APIs and Canvas. In the right order, with an interactive course.
I read all 17 of Anthropic's official skills, frontmatter included, while building my own marketplace. The real patterns of a SKILL.md that triggers, and where the docs and the code disagree: the description (and under-triggering), the 500-line rule nobody follows, scripts vs references, and the eval-loop they actually use.
Event Sourcing and CQRS explained without the jargon: what they are, the three real wins, when to use them (and when not to), the complete mental model and a Go + PostgreSQL implementation path linking the whole series.
I sent AI digging through real CV studies to score my own. Surprise: the '75% rejected by ATS' and '+40% callbacks from numbers' stats are fabricated. What actually holds up, the one real technical trap, and a radar to score a CV on what matters.
Once you grasp the OOP concepts, PHP makes them concrete: typed classes, visibility, interfaces, traits, namespaces and PSR-4 autoloading. The learning order, with runnable code.
Five automations delivered to small businesses: competitor price scraping, quote follow-ups, multi-platform stock sync, PDF reporting, lead qualification bot. Real prices and what went wrong.
OOP is understood through its concepts (object, class, encapsulation, inheritance, composition, polymorphism) before any language syntax. The method, and the 'model the real world' trap.
A page going from 80 ms to 9 seconds in prod, because of a bug-free ten-line loop. Big O explained through that real case: the O(n²) loop, the array vs linked list trap, and the Set that fixes it all. No math.
A reader asked for "the ultimate cart with all the patterns". Here is how the real ideal cart got built: five patterns each hired by a nameable need, two refused, and a deliberately dumb central class.