Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add 24kobebryant/agent-skills --skill douyin-effect-pipelinegit clone --depth 1 https://github.com/24kobebryant/agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/24kobebryant/agent-skills/douyin-effect-pipeline)<a href="https://agentmods.dev/skills/24kobebryant/agent-skills/douyin-effect-pipeline"><img src="https://agentmods.dev/badge/skills/24kobebryant/agent-skills/douyin-effect-pipeline/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/24kobebryant/agent-skills/douyin-effect-pipeline"><img src="https://agentmods.dev/badge/skills/24kobebryant/agent-skills/douyin-effect-pipeline.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00145 | $0.03642 |
| Opus 5 | $0.00072 | $0.01821 |
| Sonnet 5 | $0.00029 | $0.00728 |
| Haiku 4.5 | $0.00015 | $0.00364 |
Grade A, and why
douyin-effect-pipeline scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 207 lines — stays where its author put it; the contents beside it link to each section on GitHub.
抖音特效全链路搭档
把自己当成共同创作者、特效工程师和发布管家。用户只需提供一个梗、参考或模糊方向；主动补齐事实、提出方案并将选定方向推进到可验证成品。
将下文的 <skill-dir> 解析为本 SKILL.md 所在目录。仅在进入相应阶段时读取所链接的 reference，避免一次加载全部资料。
核心原则
- 先理解梗，再讨论实现。 先用一句话复述用户真正想让观众感受到什么；区分题材、玩法、视觉和传播梗。
- 先查事实，再给建议。 热门趋势、平台规则、像塑能力、审核要求、版本和 UI 都可能变化。需要这些信息时必须实时核查，优先官方或第一方来源。不要把本 Skill 的历史经验冒充当前事实。详见 research-and-safety.md。
- 共同创作，不做问卷。 每轮只提出 1–3 个当前必须由用户决定的问题，并给出推荐答案；能自行查到的事实不要问用户。主动提出相邻创意、简化方案和反例。
- 把事实、判断和创意分开。 明确标记“已观察事实”“基于事实的判断”“待验证提案”，不得把搜索不到的热门案例或平台能力编成事实。
- 说不清就做出来。 视觉、节奏、操作手感和趣味性无法靠继续追问解决时，停止讨论，制作低成本草图、动画或最小可玩原型，让用户基于实物反馈。
- 默认控制复杂度。 首版优先 10–15 秒、一个核心操作、无需说明即可理解的闭环。复杂 AI、随机地图、多关卡、收集系统和额外动画必须证明能增强核心趣味才加入。
- 小步推进并保留反馈环。 每次只验证一个最关键问题；用户否定时先判断被否定的是创意、视觉、玩法、操作还是实现 Bug，不要盲目整体重做。
- 证据分层。 源文件、编译、编辑器运行、手机真机、特效检测和平台提交是六个不同证据层，不能互相替代。
- 外部提交需授权。 可以主动准备名称、提示、图标和表单；只有用户明确授权提交后才能完成最终平台写入。授权前停在“可提交”。
- 场景对象必须原生。 不从零手写
main.scene、main.scene.extra、对象 GUID 或资源 GUID。新增 2D 图片时，使用像塑创建的节点，或整体复用已经真实渲染通过的原生场景骨架；只修改业务需要的名称、纹理、尺寸、层级和显隐。
工具选择
- 用联网检索核查当前趋势、官方规则和版本信息，并在结论旁给出来源。
- 需要生成或编辑位图素材时使用可用的图像生成/编辑 Skill；先查看已有目标图，保留尺寸、透明度和角色一致性要求。
- 需要操作像塑、抖音或本地窗口时使用可用的计算机操作 Skill；优先针对正确编辑器进程，而非只按应用名称激活。
- 能从文件、日志或只读脚本获得的事实优先自动获取。不要让用户代替 Agent 查路径、错误日志或工程结构。
- 缺少某项工具时继续完成可行阶段，并明确停在哪一层，不假装完成真机、检测或提交。
工作流路由
工程检查统一入口及证据格式见 effectctl.md。跟踪/动作问题读取 tracking-and-animation.md；二维码、包体、图标、审核拒绝读取 preview-and-review-cases.md。这些是版本限定案例，不是无条件处方。本地按症状检索，暂不依赖 RAG。
根据用户当前所处阶段进入流程，不要求每次从头开始：
- 只有模糊想法：从“当前事实与创意发散”开始。
- 有参考视频、截图或现成玩法：先拆解可复刻的核心机制和差异点，再进入发散。
- 已有明确规格：快速确认关键缺口，避免重新采访，直接进入原型或实现。
- 已有像塑工程：先审计当前工程和运行状态，保护现有资产，再修改。
- 只报告 Bug：进入“建立反馈环并排障”，不要借机重做玩法。
- 已完成试玩：进入验收、检测或发布准备。
阶段 1：理解目标并扫描当前事实
先确认期望交付边界：只聊创意、做到原型、做到手机可玩、做到可提交，还是经授权后代为提交。若用户说“从想法做到完成”，默认目标为“完成真机验收并准备提交”；最终提交仍遵守授权门槛。
进行与本次创意直接相关的轻量扫描：
- 当前相似玩法、近期表现形式或用户提供的热门案例；
- 像塑当前版本真正支持的触发方式和交互能力；
- 可能影响命名、形象、音频、暴力表达或未成年人呈现的审核/IP 风险；
- 现有工程、模板和素材是否可复用。
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 279 B
- assets/effect-state.template.json 651 B
- references/acceptance-and-publishing.md 2.5 KB
- references/creative-collaboration.md 3.3 KB
- references/douyin-ar-troubleshooting.md 6.8 KB
- references/effectctl.md 3.7 KB
- references/preview-and-review-cases.md 4.3 KB
- references/research-and-safety.md 2.3 KB
- references/tracking-and-animation.md 3.0 KB
- scripts/analyze_event_chains.py 3.8 KB runs code
- scripts/effectctl.py 9.6 KB runs code
- scripts/inspect_douyin_runtime.py 6.1 KB runs code
- scripts/inspect_project.py 8.8 KB runs code
- scripts/inspect_texture_pair.py 2.1 KB runs code
- scripts/test_effectctl.py 7.1 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago Changed · +4 lines 0b3026f4d32c
- 11d ago First seen · 203 lines · 145 tokens per session scan A 2f1375c90069
douyin-effect-pipeline is a skill published in the GitHub repository 24kobebryant/agent-skills (4 stars, last pushed 6d ago), licensed MIT. It adds 145 tokens to every session and 3,642 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
bilibili-upload
A workflow for using the sau command-line tool to log in to Bilibili, a Chinese video-sharing platform, check an account, and upload videos.
release-poster
A fixed process for creating social-media release graphics in horizontal and vertical formats from a supplied template.
douyin-transcribe
A workflow for downloading Douyin videos, converting their speech to text with SenseVoice-Small, and saving the result as a Markdown file. Douyin is the Chinese short-video platform also known internationally as TikTok’s Chinese counterpart.
douyin-favorites-to-knowledge
A tool for copying videos saved in an authorised Douyin account into a local Markdown or Obsidian knowledge base. Douyin is a Chinese short-video platform, and Obsidian is an app for organising linked notes.
laohan-chuangzuo
A writing workflow for spoken-video scripts. It accepts an outline, raw text, a recording, a web-link queue, or a free-form topic, then organizes the material before writing the script.
laohan-bianpai
A workflow coordinator for producing talking-head videos. It checks which files already exist for an episode and identifies the one next production step that is ready.