xyaz1313/xyskill / skills/xy-ai-workflow
xy-ai-workflow
Verified【AI流水线】内容生产 AI 流水线——先审计你现在从想法到发布每一步是谁在做、卡在哪,再判定哪一步该给 AI、哪一步必须留人,然后把想法→选题→逐字稿→切片→标题→发布排雷编排成一条有交接物的流水线,教你"某个环节 AI 跑通了就固化成 skill"的方法,并划清文风复制的正当边界(复制结构,不复制别人的稿)。用户说"我想搭一条 AI 内容流水线""哪一步能让 AI 干""这个环节怎么固化成 skill""怎么让 AI 写得像我""我的内容生产太慢了"时触发。|小爷出品
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 35.8 KB |
| agents/openai.yaml | config | 319 B |
| evals/evals.json | config | 2.3 KB |
| references/atoms.jsonl | other | 314.6 KB |
| scripts/atoms-search.pyexecutable | script | 11.5 KB |
evals/evals.json
View on GitHub[
{
"input": "我一个人做知识付费账号,现在流程是:刷到灵感记备忘录→想到就写口播稿→自己拍→剪辑→随便起个标题发。一周只能出 2 条,太慢。哪些步骤能让 AI 干?",
"expect": "先按工位表把流程摊开(想法/选题/逐字稿/拍录/切片/标题/排雷/自检/复盘),逐工位问四问(规则说得清?错的代价?1% 差距值不值?需不需要立场判断);给出人/AI 初稿人终审/AI 的分工判定,默认选题、逐字稿终稿、发布决策留人,素材提取、切片方案、标题候选、排雷初筛给 AI;定义交接物(文件)与停顿点;判定哪个环节满足固化三条件(稳定标准/重复次数/测试用例);不推荐具体 API 中转或付费工具;不复述 prompt 原文;引用原子 id;每 2 个小节停顿;结尾给'明天第一件事:把某一个工位手动跑 3 次并存下输入输出'。",
"should_route_elsewhere": false
},
{
"input": "有个博主的文案我特别喜欢,我把他 30 篇稿子存下来了,想让 AI 学他的风格给我写,怎么做?",
"expect": "进入文风边界审查:区分'像自己'与'像别人',区分借结构与借稿;明确不能把他人稿子直接喂给 AI 仿写或洗稿,可以拆其开头结构、论证顺序、受众关系与心理触发机制并模板化套自己的业务内容;建议从用户自己的历史稿抽象风格规则、用负反馈规则积累让 AI 像'我';提醒样本要同一说话逻辑分组、规则不写模板、禁'反问+感叹'说教句式;引用 XY-MC-0230 / XY-MB-152 / XY-BM-242 / XY-MB-154 类原子;不给 prompt 原文。",
"should_route_elsewhere": false
},
{
"input": "这是我明天要发的口播稿,帮我看看开头会不会被划走、有没有 AI 味、能不能直接发。",
"expect": "识别为单篇内容诊断/开头/AI 特征/发布排雷的工位级任务,不是流水线设计;说明边界,问一句'要的是这一篇,还是想让以后每一篇都省这一步';用户只要这一篇时保留稿件信息并结束,不代写不改稿不出风控结论;由 /xy 重新判断(方向上属于 xy-content-scan / xy-opener / xy-human-touch / xy-publish-guard 的职责)。",
"should_route_elsewhere": true
}
]
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |