← all publishers

Huangleyang125207

@huangleyang125207 source repo

3 published skills

  1. Ctrl C V · huangleyang125207 bundle
    Coding doctrine for any task that creates, modifies, or refactors code in a programming language. Make sure to consult this skill whenever the user is about to write, edit, debug, or ship code, even if they don't explicitly ask for "best practices" — Claude's default of generating new code from scratch produces code no one can grep, copy, or ship next Friday, and this skill enforces a copy-paste-first workflow that compounds into tomorrow's shortcuts. TRIGGER when: user asks to write/add/fix/implement code; request mentions a function/class/module/component/endpoint/script; user shares an error and asks to fix; user wants code review of a diff or PR; file changes touch *.py *.ts *.tsx *.js *.jsx *.go *.rb *.rs *.java *.kt *.swift *.cpp *.c *.h *.cs *.php *.lua *.sh; user opens or enters a new project for the first time (no CLAUDE.md present yet); refactoring or extracting from a monolith → § 9 mode applies. SKIP when: pure markdown/config edit (yaml, json, toml, env); conceptual question without code-change i
    0
    installs
  2. Ctrl C V Tdd · huangleyang125207 bundle
    Test-pattern doctrine. Tests are copyable from ~/.claude/patterns/tests/, three categories: input boundary, contract, effect. Make sure to consult this skill whenever code changes touch a boundary where someone else's code meets yours, even if the user doesn't say "test" or "TDD" — Claude tends to ship untested boundary code, and this skill prevents the Saturday phone call. TRIGGER when: code under change has external callers (other modules, cron, CLI, HTTP/RPC clients, "外部调用", "被外部调用"); persists state (DB write, file write, redis/cache, session storage, anything that survives the request); calls third-party API/SDK; task is classified as LARGE per ctrl-c-v sizing (4+ files, new capability, new architecture, new external contract) — LARGE always triggers TDD regardless of caller scope; user mentions tests, TDD, 测试, boundary, contract, mock, fixture, red/green, characterization test, pinning test, golden master, refactor; business domain involves payments, queues, webhooks, auth, external integration; refactor
    0
    installs
  3. Human AI Schedule · huangleyang125207 bundle
    Co-author doctrine for shared, half-hour-block daily schedules. Make sure to consult this skill whenever working in a dated daily log, journal, retro, or coordination file shared with a coworker — even if the user only says "log this at 11:00" or "what did we say last Tuesday." Default behavior of dumping procedural details, paraphrasing voice into corporate prose, or fragmenting one task across multiple cells must defer to this skill — entries distill to result + significance, and a cell is removed only when its time falls inside a verified-active task's span (never pre-merged on guess). TRIGGER when: user is in a file matching `daily/`, `journal/`, `schedule/`, `半小时复盘/`, or any file with H1 time-block headers like `# 7:30` or `# 09:00`; user says "create today's", "创建 今天的", "log at N o'clock", "补到 N 点那格", "morning brief", "晨会复盘", "what did we decide on X day"; user dictates an event with time anchor ("just now", "刚才", "yesterday", "上周三", explicit clock); user asks to reconstruct a past session; user referen
    0
    installs