Skip to content

ZJU-REAL/Easel / skills/builtin/skill-cross-platform-publish

skill-cross-platform-publish

Verified

跨平台一键发布:一份内容适配并发布到多个平台(小红书/抖音/B站/公众号/快手/视频号/知乎/今日头条)。 按各平台格式约束(字数/比例/标签/内容类型)适配内容,再逐个委派对应平台发布 SKILL。 当用户说"一键发布""同时发到多个平台""多平台发布""一稿多发""全平台发""发到抖音+小红书+B站" "分发到各平台"时使用。适配由 LLM 做,路由与约束检查由 scripts/publish_dispatch.py。

FileKindSize
EASEL-META.mdmarkdown588 B
SKILL.mdskill md4.4 KB
scripts/publish_dispatch.pyscript9.0 KB

scripts/publish_dispatch.py

View on GitHub
#!/usr/bin/env python3
"""publish_dispatch.py — 跨平台一键发布的派发规划(平台约束注册表 + 派发计划)。

一份内容 → 多平台发布。本脚本提供:各平台的格式约束(字数/比例/标签上限/内容类型)、
平台→发布 SKILL 的路由、以及"派发计划"生成(校验内容是否越限、给出每个平台该调哪个 publisher)。

实际的按平台改写(标题/正文/标签本地化)由 LLM 完成;实际发布由各平台 publisher SKILL 执行。
本脚本是确定性的"约束检查 + 路由 + 计划",可离线测试。

子命令:
    platforms  列出支持的平台及约束
    plan       读内容清单 → 输出每平台派发计划(校验越限 + 路由到 publisher)
    selftest   自检

用法举例:
    publish_dispatch.py platforms
    publish_dispatch.py plan --manifest content.json
    echo '{...}' | publish_dispatch.py plan --manifest -
"""
from __future__ import annotations

import argparse
import json
import sys

# 平台注册表:约束 + 对应发布 SKILL
# title/body 上限为字符数;0 表示无硬限或不适用。aspect 为推荐视频比例。
PLATFORMS: dict[str, dict] = {
    "xiaohongshu": {"publisher": "skill-xhs-publisher", "types": ["image", "video"],
                    "title": 20, "body": 1000, "tags": 10, "aspect": "3:4/9:16",
                    "note": "标题≤20字;图 3:4 或 1:1,视频竖版;话题#放正文"},
    "douyin": {"publisher": "skill-douyin-upload", "types": ["video", "image"],
               "title": 30, "body": 1000, "tags": 5, "aspect": "9:16",
               "note": "作品标题≤30字(发布脚本硬限);文案/简介另计;竖版 9:16;话题# 3-5 个"},
    "wechat": {"publisher": "skill-wechat-publisher", "types": ["article"],
               "title": 64, "body": 0, "tags": 0, "aspect": "-",
               "note": "公众号图文;首图 2.35:1;正文富文本"},
    "bilibili": {"publisher": "skill-bilibili-upload", "types": ["video"],
                 "title": 80, "body": 2000, "tags": 10, "aspect": "16:9",
                 "note": "标题≤80字;横版 16:9 为主;分区+标签必填"},
    "kuaishou": {"publisher": "skill-kuaishou-upload", "types": ["video", "image"],
                 "title": 500, "body": 500, "tags": 5, "aspect": "9:16",
                 "note": "竖版 9:16;描述含话题#"},
    "weixin-channels": {"publisher": "skill-channels-upload", "types": ["video"],
                        "title": 22, "body": 1000, "tags": 5, "aspect": "9:16",
                        "note": "视频号;标题短≤22字;竖版;可带话题与位置"},
    "zhihu": {"publisher": "skill-zhihu-publisher", "types": ["article", "answer"],
              "title": 100, "body": 0, "tags": 5, "aspect": "-",
              "note": "文章/回答;专业排版;话题标签"},
    "toutiao": {"publisher": "skill-toutiao-publisher", "types": ["article"],
                "title": 30, "body": 0, "tags": 5, "aspect": "-",
                "note": "头条号图文;标题≤30字;资讯式长文 800-2500 字;话题≤5 个;可存草稿"},
}


def _die(msg: str, code: int = 1) -> None:
    print(f"ERROR: {msg}", file=sys.stderr)
    sys.exit(code)


def cmd_platforms(_a) -> int:
    print(f"支持 {len(PLATFORMS)} 个平台:\n")
    for name, c in PLATFORMS.items():
        tl = f"标题≤{c['title']}" if c["title"] else "标题无限"
        print(f"  {name:16s} → {c['publisher']}")
        print(f"    类型 {'/'.join(c['types'])} · {tl} · 标签≤{c['tags']} · 比例 {c['aspect']}")
        print(f"    {c['note']}")
    return 0


def _check_platform(name: str, content: dict) -> dict:
    """对单平台做约束检查,返回派发条目(含 warnings)。"""
    if name not in PLATFORMS:
        return {"platform": name, "ok": False,
                "error": f"未知平台(支持:{', '.join(PLATFORMS)})"}
    c = PLATFORMS[name]
    warns: list[str] = []
    title = content.get("title", "") or ""
    body = content.get("body", "") or ""
    tags = content.get("tags", []) or []
    mtype = content.get("media_type", "")

    if c["title"] and len(title) > c["title"]:
        warns.append(f"标题 {len(title)}>{c['title']} 字,需精简/截断")
    if c["body"] and len(body) > c["body"]:
        warns.append(f"正文 {len(body)}>{c['body']} 字,需精简")
    if c["tags"] and len(tags) > c["tags"]:
        warns.append(f"标签 {len(tags)}>{c['tags']} 个,需删减")
    if mtype and mtype not in c["types"]:
        warns.append(f"内容类型 {mtype} 与该平台支持 {c['types']} 不符(需换形式或转制)")

    return {"platform": name, "ok": True, "publisher": c["publisher"],
            "recommend_aspect": c["aspect"], "constraints_note": c["note"],
            "warnings": warns,
            "action": f"按约束适配后调用 {c['publisher']} 发布"}


def cmd_plan(a) -> int:
    raw = sys.stdin.read() if a.manifest == "-" else _read(a.manifest)
    try:
        m = json.loads(raw)
    except json.JSONDecodeError as e:
        _die(f"manifest 不是有效 JSON:{e}")
    content = m.get("content", {})
    platforms = m.get("platforms", [])
    if not platforms:
        _die("manifest.platforms 为空(需目标平台列表)")
    plan = [_check_platform(p, content) for p in platforms]
    out = {"content_title": content.get("title", ""),
           "target_count": len(platforms),
           "dispatch": plan,
           "hint": "对每个平台:按 constraints_note 用 LLM 适配标题/正文/标签/比例,"
                   "再调用 publisher 发布;发布后可接 skill-publish-notify / skill-publish-log。"}
    print(json.dumps(out, ensure_ascii=False, indent=2))
    return 0


def _read(path: str) -> str:
    from pathlib import Path
    p = Path(path).expanduser()
    if not p.is_file():
        _die(f"manifest 不存在:{p}")
    return p.read_text(encoding="utf-8")


def cmd_selftest(_a) -> int:
    print("publish_dispatch 自检 ...", file=sys.stderr)
    # 路由映射存在
    assert PLATFORMS["xiaohongshu"]["publisher"] == "skill-xhs-publisher"
    assert PLATFORMS["toutiao"]["publisher"] == "skill-toutiao-publisher"
    # 头条标题上限 30 字:31 字应告警,30 字不应告警
    over = _check_platform("toutiao", {"title": "标" * 31, "media_type": "article"})
    assert any("标题" in w for w in over["warnings"]), "头条标题超长未告警"
    fit = _check_platform("toutiao", {"title": "标" * 30, "tags": ["a"],
                                      "media_type": "article"})
    assert not fit["warnings"], f"头条合规内容不应告警:{fit}"
    # 约束检查:超长标题应告警
    r = _check_platform("xiaohongshu", {"title": "一" * 30, "tags": list(range(12)),
                                        "media_type": "article"})
    assert r["ok"] and r["warnings"], "越限内容应产生 warnings"
    assert any("标题" in w for w in r["warnings"]), "标题超长未告警"
    assert any("标签" in w for w in r["warnings"]), "标签超量未告警"
    assert any("类型" in w for w in r["warnings"]), "类型不符未告警"
    # 合规内容无告警
    ok = _check_platform("douyin", {"title": "短标题", "tags": ["a", "b"],
                                    "media_type": "video"})
    assert ok["ok"] and not ok["warnings"], f"合规内容不应告警:{ok}"
    # 未知平台
    bad = _check_platform("nope", {})
    assert not bad["ok"], "未知平台应 ok=False"
    # plan 端到端(stdin 风格)
    import io
    m = {"content": {"title": "测试", "body": "正文", "tags": ["x"], "media_type": "video"},
         "platforms": ["douyin", "bilibili", "kuaishou"]}
    import tempfile
    from pathlib import Path
    with tempfile.TemporaryDirectory() as td:
        f = Path(td) / "m.json"
        f.write_text(json.dumps(m), encoding="utf-8")
        # 直接调用 _check_platform 汇总(cmd_plan 打印到 stdout)
        plan = [_check_platform(p, m["content"]) for p in m["platforms"]]
        assert len(plan) == 3 and all(x["ok"] for x in plan), "plan 生成异常"
        assert plan[0]["publisher"] == "skill-douyin-upload"
    print("✅ selftest 通过(约束检查/告警/路由/plan)")
    return 0


def main() -> int:
    ap = argparse.ArgumentParser(
        description="跨平台一键发布派发规划(平台约束注册表 + 路由 + 计划)",
        formatter_class=argparse.RawDescriptionHelpFormatter)
    sub = ap.add_subparsers(dest="cmd")

    sub.add_parser("platforms", help="列出支持平台及约束").set_defaults(func=cmd_platforms)

    p = sub.add_parser("plan", help="内容清单 → 派发计划")
    p.add_argument("--manifest", required=True, help="内容清单 JSON 路径(- 为 stdin)")
    p.set_defaults(func=cmd_plan)

    sub.add_parser("selftest", help="自检").set_defaults(func=cmd_selftest)

    a = ap.parse_args()
    if not getattr(a, "func", None):
        ap.print_help()
        return 1
    return a.func(a)


if __name__ == "__main__":
    sys.exit(main())
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: 472kUpdated Claude CodeCodexVerifiedInstall
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: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
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 availableStars: 322kUpdated Claude CodeCodexVerifiedInstall