Hands-on prompt engineering and LLM evaluation. Built around a diagnose-then-ship loop: forensic transcript critique ▎ that names the exact failure mode, paired with a fixer that produces complete, deployment-ready rewrites — not ▎ vague advice.
한글 문장 한 줄만 쓰면 Playwright E2E 테스트 spec, 실행 명령어, 예상 오류, Fix Prompt까지 한 번에 만들어주는 테스트 자동화 스킬입니다. 프론트/풀스택 개발자와 소규모 QA 팀이 한국어 기획을 테스트 코드로 옮기는 시간을 줄여줍니다.
Most broken skills are not broken. They load, they validate, and they are never invoked - because the description does not match how anyone actually asks. Covers the one-character YAML mistake that makes frontmatter load empty with no error, how to write a description as a match target rather than a summary, and how to prove invocation in two steps instead of hoping.
Autopsy and rewrite the first second of your Shorts — eight field-tested failure patterns, ranked rewrites, and a promise ledger the body must pay.
Before you write code, kill the causes of rework. A phase-gate skill that runs a pre-implementation diagnosis (clarify requirements, success criteria, non-goals, hidden risks) and sets up the dev environment (toolchain check, repo init, dependency/secret/config templates, DB/API contract, folder layout). Each gate must pass before moving on. Delivers a diagnosis report, a stack-selection doc, and a ready-to-build foundation to hand off to implementation.
Point it at your web repo and it scaffolds a complete Playwright E2E UI-test suite for you — auto-detects your router (Nuxt / Next / Vue Router / React Router), discovers your routes, plans the data-testid tagging, and generates ready-to-run mock-backed and real-backend specs, a playwright.config, and a login setup project. Zero config.
Pre-submission quality gate for AI-generated broadcast scripts. Catches timing errors, sponsor name mistakes, host voice drift, missing stage directions, and unverified facts before they reach production. 8 documented failure patterns. Works for podcasts, YouTube, and live broadcast.
A kickoff protocol (agent skill) for multi-step tasks. Before writing any code, the agent must: run a spec interview, batch every human-authorization step into one checklist, build a self-verification loop (log sink + verify script + dry-run) *before* the feature itself, and build comparable variants instead of blocking on third-party decisions.
Get to the root of why loadbearing instructions fail You get the Drop Table, a plain-English root cause, a rebuild, and a scoreboard you keep