andyhuo520/vibe-knowledge / skills/ai-teardown
ai-teardown
Unverified「AI 拆机课」系列课工作室:输入一个主题(AI 产品、开源项目/书、人物历史、软件教程),产出由浅入深、每句话有出处的中文讲解视频系列 + 每节互动版 + 课程总页。流水线:收集资料 → 确认对象与课时 → 大纲 → 逐节旁白(MiMo 克隆音色 + 逐字对齐)→ scenes.json 画面编排(Notion 极简手绘引擎 / 马卡龙模板引擎,IconPark 图标与插画、Noto 动态 emoji、AntV X6 流程图、像素讲解小人、截图圈注、终端录屏)或 film.js 代码成片(cine 引擎:Three.js、粒子、着色器、真片段叠加、曲库配乐)→ Hyperframes / 无头 GPU 逐帧 1080p 渲染 → Artifact 发布 → 课程总页 + 导学视频。触发:做系列课 / 做课程视频 / 把某本书或某个开源项目做成课 / 科普讲解视频 / 教学视频 / 课程地图 / 导学视频 / 互动课件 / 「拆机课」「上手课」「实战课」「编排课」/ 王炸讲解片 / 用真实会话和成片教人做视频。不用于:单支宣传片(用 promo-video)、单张信息图、纯文章写作。
engine/scripts/check_scenes.py
View on GitHub#!/usr/bin/env python3
"""画面编排自检:scenes.json 里每个时间引用(B03 / B03.cue / B03.end,可带 ±秒)都要在 narration.json 里有对应的拍和 [[cue]]。
不用等配音,写完 scenes.json 就能跑。用法:python3 scripts/check_scenes.py <lessonDir>…"""
import json, re, sys
from pathlib import Path
REF = re.compile(r"^(B\d\d)(?:\.([a-z0-9_]+(?:-[a-z][a-z0-9_]*)*))?(?:@[^+-]+)?(?:[+-][\d.]+)?$")
def walk(o, path=""):
if isinstance(o, dict):
for k, v in o.items():
if isinstance(v, str) and (k in ("at", "settle", "hi", "celebrateAt", "titleAt", "questionAt", "eyebrowAt", "peopleOut", "drawnBy", "vsAt", "chatAt", "iconAt", "srcAt", "boxAt", "out") or k.endswith("At")):
yield f"{path}.{k}", v
else:
yield from walk(v, f"{path}.{k}")
elif isinstance(o, list):
for i, v in enumerate(o):
yield from walk(v, f"{path}[{i}]")
ROOT = Path(__file__).resolve().parent.parent
SETS = {}
for f in (ROOT / "assets/iconsets").glob("*.json"):
j = json.load(open(f)); SETS[j.get("prefix", f.stem)] = set(j["icons"]) | set(j.get("aliases", {}))
def icons(o):
if isinstance(o, (dict, list)):
for v in (o.values() if isinstance(o, dict) else o):
yield from icons(v)
elif isinstance(o, str) and re.match(r"^[a-z-]+:[a-z0-9-]+$", o): yield o
bad = 0
for d in map(Path, sys.argv[1:]):
beats = {b["id"]: set(re.findall(r"\[\[([a-z0-9_-]+)\]\]", b["narration"])) for b in json.load(open(d / "narration.json"))["beats"]}
spec = json.load(open(d / "scenes.json"))
order = []
for p, v in walk(spec):
m = REF.match(v)
if not m: print(f"{d.name} {p}: 看不懂 {v!r}"); bad += 1; continue
b, cue = m.groups()
if b not in beats: print(f"{d.name} {p}: 没有这一拍 {b}"); bad += 1
elif cue and cue not in ("end", "cue") and cue not in beats[b]: print(f"{d.name} {p}: {b} 里没有 [[{cue}]](有:{' '.join(sorted(beats[b]))})"); bad += 1
for ic in set(icons(spec)):
pre, name = ic.split(":")
if pre in SETS and name not in SETS[pre]: print(f"{d.name}: 图标库里没有 {ic}"); bad += 1
tj = d / "build" / "timings.json"
if tj.exists(): # 有了配音时间,再查一遍:场景太短,观众来不及看
TJ = json.load(open(tj)); T = TJ["beats"]; SPD = TJ.get("speed", 1)
T = T if isinstance(T, dict) else {b["id"]: b for b in T}
def when(r):
m = re.match(r"^(B\d\d)(?:\.([a-z0-9_]+(?:-[a-z][a-z0-9_]*)*))?(?:@([^+-]+))?([+-][\d.]+)?$", r)
b, cue, word, off = m.groups(); B = T[b]
if word: # 「B03@机器」:读到这个词的那一刻(与 lesson.js 的 W() 一致)
i = B["text"].find(word)
base = B["chars"][i] if i >= 0 and B.get("chars") else B["start"]
if i < 0: print(f"{d.name}: {b} 旁白里没有「{word}」")
else:
base = B["end"] if cue == "end" else B["cues"].get(cue, B["start"]) if cue else B["start"]
return base + float(off or 0) / SPD
ts = [when(sc["at"]) for sc in spec["scenes"]] + [max(B["end"] for B in T.values())]
for k, sc in enumerate(spec["scenes"]):
if ts[k + 1] - ts[k] < 2.5: print(f"{d.name}: 第 {k} 个场景({sc['t']} @ {sc['at']})只有 {ts[k+1]-ts[k]:.1f} 秒"); bad += 1
if ts[k + 1] < ts[k]: print(f"{d.name}: 第 {k + 1} 个场景比前一个还早"); bad += 1
starts = [s["at"] for s in spec["scenes"]]
print(f"{d.name}: {len(spec['scenes'])} 个场景,覆盖 {len({s.split('.')[0][:3] for s in starts})}/{len(beats)} 拍的起点")
sys.exit(1 if bad else 0)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |