laohan-fengmianzhuzige skillA
laohan-fengmianzhuzige is agent-read markdown (skill) from hanzhcn/laohan-skills: 完全对标柱子哥TzFilm封面体系（81张真实封面逐张检视实证2026-09-20）——三层合成管线：AI/真实背景板（无字无人）+真人环境照片层+Remotion材质级文字层。画幅1080x1440(3:4原生)。Use when 工作流⑥选用zhuzige套（默认），或用户说"柱子哥封面""做封面"。与 laohan-fengmianqiuzhi（秋芝套）互为可切换双轨。.
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# 柱子哥式封面：三层合成 v6.1（定版） **v6.1定版（Jeffrey验收2026-09-20）**：`samples/shoucuogongju-v62-final.jpg`（SHA-256前缀`aeb933ed`）为风格基线样张——OPPO形象照（藏青衬衫+眼镜+胡茬）+全栈v6.2（3:4材质字+左中证据卡EvidenceCard+右上MiniStamp小章+人物提亮档）。定版参数基线=samples/brief-example.json：人物提亮档（env_brightness 0.66/saturation 0.78/tint 0.26/face_boost 1.7/弱scrim 160）+text_top 748+kicker 236+证据卡三行（内容取自当期口播稿深水区清单）。同日实测教训：暗场单侧光素材（微信头像照）若脸的一半在暗部+裁切不当会"脸看不全"——**素材验收第一优先级=脸完整明亮可见**，其次才是光比融入；定版因此回到亮场形象照+提亮档路线。出厂检查纪律：三段放大逐字转录+关键边界像素级检测（章×脸间隙、脸区零遮挡、无全黑行），不得只靠整体打分。 **路线定义**（81张真实封面逐张检视实证 + 2026-09-20复刻实测）：柱子哥封面**不是整幅AI生成**，是三层合成——①AI生成/UI截图的**无字背景物件板**；②他本人**真实照片**（每期服装姿势不同，脸100%一致靠真人）；③设计工具排版的**五层文字系统**；合成后**整图统一调色**（冷调暗调ONE GRADE——先合成后调色，人物与环境才是一张照片）。2026-01他本人试过"AI生成整幅含人物"（#03/#04）后放弃，2026-03起全部转为三层合成。本skill用同一结构复刻。 **历史教训链**：v2.0误判"纯抽帧"（抓错图层）→ v3.0误改"AI单pass生成整幅海报"（身份不像/英文错字ERLY/版式失控三硬伤实测）→ v4.0三层合成确立 → v4.x小窗/抠像"贴纸感"翻车 → v5.1定案`env_photo`唯一正解 → **v6.0实测钉死三个非参数级瓶颈并逐项修掉**：(a)画幅错位——参考系81张全部3:4(covers_real 323x430)而产物曾是9:16，画幅改为1080x1440原生；(b)文字材质——PIL平面填充=PPT感，改为**Remotion静帧渲染材质字**（与正片HUD同源组件，templates-v2已验收）；(c)素材光线——亮场平光照片压暗后=灰暗监控感（v6加入机器可判的亮场预检），**暗场摆拍是env_photo路线的硬前提**。 ## 输入与前置 1. 当前 `01-口播稿.md`：取hook句+主记忆点→转标题组； …
Read the whole file at its exact version.
How to install
mdr add hanzhcn/laohan-skills/laohan-fengmianzhuzige@v6.1.0mdr add hanzhcn/laohan-skills/laohan-fengmianzhuzige@sha256:63b1a71c7363697fPin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_xam7eoh42awi26jf)
1 badge views in 30 days
Versions
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v6.1.0 latest | 2026-09-20 | 9e742bd | 8,655 B | A | view · diff |
| v6.0.0 | 2026-09-20 | 55ac907 | 7,712 B | A | view · diff |
| v5.1.0 | 2026-09-19 | 751f30d | 6,051 B | A | view · diff |
| v5.0.0 | 2026-09-19 | 0889afc | 5,473 B | A | view · diff |
| v4.1.0 | 2026-09-19 | fb23e77 | 4,671 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (8655 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
hanzhcn/laohan-skills · 11 stars · license MIT · pushed 2026-09-20 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_xam7eoh42awi26jf GET https://markdownregistry.com/api/v1/resolve?ref=hanzhcn/laohan-skills/laohan-fengmianzhuzige GET https://markdownregistry.com/api/v1/blob/63b1a71c7363697f806966c1b67de2a90e0794007e805b8059a40d422cb4b63d
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.