Agent skill

Imagegen

by yan-labs in yan-labs/yan-skills

当用户要求生成图片、配图、插图、logo、吉祥物、封面、海报、og 图、favicon 源图、真人场景图,或建站时页面缺图、仍有占位图时使用。包括“画一张”“出一套图”“网站需要配图”“image gen”。本 Skill 负责提示词、生成、三道验收、压缩与落盘;只是让 Codex CLI 做代码代理任务时用 codex,普通派单走 agent-fleet,页面 SEO 与上线闸门走…

MITAuto-check passedMedia & Creative

Install Imagegen

skills CLI
$ npx skills add yan-labs/yan-skills --skill imagegen -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install yan-labs/yan-skills imagegen --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/yan-labs/yan-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/imagegen .claude/skills/imagegen && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
imagegen
GitHub stars
208
Token cost
~2.5k tokens
SKILL.md length
625 words
Files
3
Skills in repo
7
Repo updated
First seen
Licence
MIT

At a glance

当用户要求生成图片、配图、插图、logo、吉祥物、封面、海报、og 图、favicon 源图、真人场景图,或建站时页面缺图、仍有占位图时使用。包括“画一张”“出一套图”“网站需要配图”“image gen”。本 Skill 负责提示词、生成、三道验收、压缩与落盘;只是让 Codex CLI 做代码代理任务时用 codex,普通派单走 agent-fleet,页面 SEO 与上线闸门走…

  • Works in 6 steps: 输出目录绝对路径(已 mkdir -p)。 → 逐张编号:精确文件名 + 具体描述 + 像素尺寸/比例。 → 共享风格块:画风、背景、hex 调色板、光线、镜头——一套图靠它保持一致。 → …
  • Tasks that involve Image generation
  • SKILL.md covers 启动命令, 提示词怎么写, 按用途的模板 and 一套图的一致性, plus 6 more sections
  • Calls codex, magick and brew

What it does

Imagegen is an agent skill from yan-labs/yan-skills. 当用户要求生成图片、配图、插图、logo、吉祥物、封面、海报、og 图、favicon 源图、真人场景图,或建站时页面缺图、仍有占位图时使用。包括“画一张”“出一套图”“网站需要配图”“image gen”。本 Skill 负责提示词、生成、三道验收、压缩与落盘;只是让 Codex CLI 做代码代理任务时用 codex,普通派单走 agent-fleet,页面 SEO 与上线闸门走 rankup。

Its SKILL.md is about 2.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files (for example `README.md` and `evals/evals.json`).

It sits in Media & Creative, covering Image generation. The repository describes itself as: Yan's agent skills collection — Google Trends SEO workflows, AI news, autopilot, and more. For Claude Code / Codex / Cursor. The licence is MIT.

When your agent uses it

  • Tasks that involve Image generation

Example prompts

  • “网站需要配图”
  • “image gen”
  • “/imagegen”

Requirements

  • Python 3

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. 输出目录绝对路径(已 mkdir -p)。
  2. 逐张编号:精确文件名 + 具体描述 + 像素尺寸/比例。
  3. 共享风格块:画风、背景、hex 调色板、光线、镜头——一套图靠它保持一致。
  4. No text, no letters, no logos, no watermarks. 生成的字几乎必花,非英文界面更是错字;文字后期用 HTML/CSS 叠。
  5. 逃生口:If you genuinely cannot generate images, say so plainly. Do not substitute placeholders, ASCII art, solid rectangles, or images…
  6. 回报要求:每个文件的绝对路径、实际像素、字节数、有无 alpha、所用方法(哪个工具/模型、有无本地后处理)。

What it can do on your machine

Read from SKILL.md and the folder at commit 8ec2915. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • codex
    • magick
    • brew
    • curl

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use curl, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Imagegen loads about 2.5k tokens when it runs. Until then it costs about 53 tokens; SKILL.md has 625 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~53
When it runs · the whole SKILL.md, loaded when a task matches
~2.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from yan-labs/yan-skills at commit 8ec2915, republished under its MIT licence (© yan-labs). 625 words, ~2,477 tokens.

Download SKILL.mdSave it as .claude/skills/imagegen/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
imagegen
description
当用户要求生成图片、配图、插图、logo、吉祥物、封面、海报、og 图、favicon 源图、真人场景图,或建站时页面缺图、仍有占位图时使用。包括“画一张”“出一套图”“网站需要配图”“image gen”。本 Skill 负责提示词、生成、三道验收、压缩与落盘;只是让 Codex CLI 做代码代理任务时用 codex,普通派单走 agent-fleet,页面 SEO 与上线闸门走 rankup。
metadata.version
1.0.0

imagegen

一句话定位:借 Codex agent 内置的图像生成能力出图;本 Skill 只管五件事——把需求翻成好提示词、把命令跑通、验收、压缩、放进项目该在的目录。

图像生成不是 CLI 子命令(没有 codex image;codex exec --image 是把图当输入附上)。它是 Codex agent 的内置工具 image_gen。不要翻 codex --help 找 flag,找不到就下结论"不能生图"——这个结论是错的。描述你要的图,让 agent 自己选方法。

启动命令

bash
mkdir -p <outdir>                                # 先建目录,提示词里写它的绝对路径
cat <outdir>/prompt.md | codex exec --skip-git-repo-check \
  --config model_reasoning_effort="medium" \
  --sandbox danger-full-access \
  -C <outdir> -o <outdir>/final.md 2>/dev/null
要点说明
后台跑Bash 的 run_in_background: true,两张图约 2–3 分钟,成套图更久;等完成通知,不要轮询。注意:必须从主线程(而非 subagent)调用——subagent 用 run_in_background 后 park 等通知,会被 harness 判定空闲并终止,通知永远送达不了
--sandbox danger-full-access生图要走网络。这个 flag 是否需要确认取决于当前机器的授权设置:有常设授权就直接跑,没有就按该机器的规则确认一次;无论哪种,启动那一行都要说明用了哪个 sandbox
-o <outdir>/final.md最终报告写进文件,从这里读路径与方法;stdout 是进度噪音,别去解析
2>/dev/null压掉 stderr 的思考流;调试 Codex 本身时才拿掉
effortmedium 够用,这不是推理任务
不传 -m用 ~/.codex/config.toml 的默认模型,用户点名才覆盖
原图在哪Codex 每次 image_gen 的原始输出都落在 ~/.codex/generated_images/<session-id>/exec-*.png(0.7–1.5 MB/张);输出目录里的是它后处理过的版本,要原图去那里取

codex --version 失败或启动就退出:先 codex doctor,如实报告,不要盲目重试。

提示词怎么写

一份 prompt.md 必含六样,缺一样就会出一类问题:

  1. 输出目录绝对路径(已 mkdir -p)。
  2. 逐张编号:精确文件名 + 具体描述 + 像素尺寸/比例。
  3. 共享风格块:画风、背景、hex 调色板、光线、镜头——一套图靠它保持一致。
  4. No text, no letters, no logos, no watermarks. 生成的字几乎必花,非英文界面更是错字;文字后期用 HTML/CSS 叠。
  5. 逃生口:If you genuinely cannot generate images, say so plainly. Do not substitute placeholders, ASCII art, solid rectangles, or images downloaded from the web.
  6. 回报要求:每个文件的绝对路径、实际像素、字节数、有无 alpha、所用方法(哪个工具/模型、有无本地后处理)。

透明图务必写 true alpha, not a white/dark square——实测 Codex 靠这句自检,第一版烤了底色又自己返工。

再加一条常用兜底:If the exact size is unsupported, generate the nearest aspect and resize locally (sips / PIL) to the exact pixels; keep alpha for transparent items.

完整示例(吉祥物 + og:image 两张一套)
markdown
Generate two images with your built-in image generation tool and save them to:
<outdir absolute path>/

Shared style: palette #2563EB (primary) #F59E0B (accent) #0F172A (ink) #F8FAFC (paper).
No text, no letters, no logos, no watermarks anywhere.

1. `mascot-logo.png` — 1024x1024, transparent background PNG (true alpha, not a white square).
   A friendly round owl holding a tiny wrench. Flat vector style, bold clean shapes,
   2-3 flat tones per element, no gradients, no drop shadows. Occupies 75-80% of the
   frame, centered, exclude the wrench from that measure.

2. `og-image.png` — 1200x630 (1.91:1), opaque. Photorealistic editorial photo: a person
   in their late twenties at a wooden desk in front of a laptop, soft window light from
   the left, shallow depth of field, calm home office, mug and small plant. Eye-level,
   subject on the left so the right third is clean negative space. Laptop screen is a
   soft blurred blue-white glow, nothing readable.

Rules: if you genuinely cannot generate images, say so plainly — no placeholders, ASCII
art, solid rectangles, or web downloads. If the exact size is unsupported, generate the
nearest aspect and resize locally (sips / PIL); keep alpha for the mascot.
Report per file: absolute path, actual pixels, bytes, alpha yes/no, method used.

按用途的模板

用途尺寸 / 格式风格块要点必写约束
logo / favicon 源图1024×1024 透明 PNG;后续 sips -Z 缩出 512/192/180/32/16flat vector, bold silhouette, 2–3 tones, no gradients主体占 75–80%;16px 下还认得出(单一形状,不靠细节);不要文字
吉祥物1024×1024 透明 PNG;成套则同一提示词多姿势同上 + 头身比、眼径÷头宽等数字化特征物种、道具(只准一个)、配色全部写死;道具不计入占比
og:image / 分享图1200×630,不透明;每页一张独立主体偏一侧,另一侧留干净负空间(标题后期叠)缩到 120px 仍看得出主体;不要文字;不得全站共用
内页配图1600×900 或 4:3;JPEG/WebP与站点调色板一致;同一站同一画风与该页目标词语义相关(写出场景,不写关键词)
用户场景 / 真人图1600×1067(3:2)或 1200×630photorealistic editorial, natural window light, shallow DoF, 35–50mm 视角写年龄段、动作、环境、光向、构图;屏幕内容"模糊发光,无可读文字";不写真人姓名
手绘 / 蜡笔 / 水彩插画1600×1200 或方图;带纸纹则不透明crayon on textured paper / loose watercolor, visible strokes, limited palette写纸色与颗粒感;线条粗细;留白比例
海报 / 电影感画面2:3(1000×1500)竖 或 21:9 横cinematic, anamorphic, volumetric light, color grade 写成 hex 两色文字区留空;主体位置写清(三分法哪一格)

尺寸比例写清,模型给不了精确像素时靠本地缩放,见「压缩与落盘」。

一套图的一致性

  • 一个提示词文件出一套,共享风格块放最上面,逐张只写差异;分批跑时把同一段风格块原样复制。
  • 调色板用 hex 写死,不写"蓝色系";同一套里角色比例、线宽、纸纹也写成数字。
  • 形容词见顶("再日式一点")就改用可测量参数:头身比、眼径÷头宽、线宽÷图宽、HSV 饱和度区间、留白占比。把参考组和产出用同一段脚本量一遍,出「参数 | 参考区间 | 我们的值 | 判定」表,感觉才能变成可逐条修的清单。
  • 可以默认用目标语言写提示词(n=3 盲评倾向目标语言,但未达显著),真正确定有效的是把视觉约束写成数字。

验收:三道检查,缺一道漏一类问题

出处:codex/references/image-experiments.md「一套图的验收」,2026-08-22 一次 16 张角色图的实录,三道各自抓到了不同缺陷。

先用 Read 逐张打开看过,再谈下面三道;没看过的图不准接进页面。三道都是必做,缺一道漏一类问题。

  1. 接触印相(抓构图失衡):全部缩到 120px 横向拼一张——这就是结果页/分享卡的真实尺寸。全尺寸下漂亮、缩略图只剩一把椅子的图,只有这一步能暴露。 sips -Z 120 in.png --out thumbs/in.png
  2. alpha 包围盒占比(把"够不够大"变成数字):
    python
    from PIL import Image
    im = Image.open(f).convert('RGBA'); bb = im.getchannel('A').getbbox()
    frac = (bb[2]-bb[0])*(bb[3]-bb[1]) / (im.width*im.height)   # 目标 0.75–0.80
    实测一组六张 41%–66%,没有一张达标且相差 1.6 倍,并排看只觉得"有点乱"。提示词里除了写占比还要写明道具不计入。
  3. 独立盲评(抓风格与规则遵从,最容易被省掉):成对结果打乱成 pairN-A/B,对照表放项目目录之外,派一个没参与生成的 agent 评,明说「看不出差别」可接受。非盲判断曾被盲评整组反转,机制是多出来的道具。

参考图 / mood board

风格不受版权保护,喂参考图是设计行业的常规做法。控制点在输出端,不在输入端:

  1. 参考图用一组(10 张以上不同来源拼 mood board),不用单张——单张最容易长得像原图。
  2. 参考图只传质感(线条、上色、头身比),主体由我们写死:物种、道具、姿势、配色。
  3. 出图后与参考组并排做相似性检查:"这张会被认成某个已有角色吗"——像了就改主体特征重生成,不改风格。
  4. 提示词里不点名受版权保护的角色或"in the style of X":参考图已把信息传到,点名只加风险不加效果。
Show full SKILL.md (301 more words)Show less

压缩与落盘

原图约 1 MB/张 PNG,不压缩不许进仓库。

bash
sips -s format jpeg -s formatOptions 82 in.png --out out.jpg   # 不透明图:JPEG,4 MB 一套压到 1 MB 内
cwebp -q 82 in.png -o out.webp                                  # 页内用 WebP(透明也保留)
cwebp -q 82 -alpha_q 100 mascot.png -o mascot.webp              # 透明主体,alpha 无损
sips -Z 512 logo.png --out icon-512.png                         # 等比缩最长边
sips -z 630 1200 og.png --out og.png                            # 精确到像素(先高后宽)

透明没做出来、或主体占比不达标时,用 ImageMagick 二次处理(先 which magick,没有就 brew install imagemagick):

bash
magick in.png -fuzz 8% -transparent '#0F172A' out.png          # 纯色底抠透明(换成实际底色 hex)
magick in.png -trim +repage -resize 800x800 -gravity center \
  -background none -extent 1024x1024 out.png                    # 裁掉空边,主体缩到约 78% 再居中回 1024²
magick out.png -format 'alpha_min=%[fx:minima.a] colors=%k\n' info:   # alpha_min 必须是 0 才算真透明
素材保留格式放哪(示例路径)
logo / 图标集透明 PNG 源 + SVG(如有);导出 512/192/180/32/16<project>/public/brand/,manifest.json 逐个真实引用
og:imagePNG 或 JPEG(WhatsApp/FB 预览爬虫对 WebP 不稳;rankup/references/seo-experiences.md 2026-07-18)<project>/public/og/<page-slug>.png,一页一张
页内配图 / 场景图WebP(同图 175 KB→56 KB),<img> 写真实 width/height<project>/public/images/<section>/
原始生成物与 prompt.md原样留档,不进 public/<project>/design/imagegen/<batch>/
  • 多文件循环、等待轮询一律用 python3 heredoc 驱动:Claude Code 的 Bash tool 里裸 for f in *.png; do …; done 和 until [ -f x ]; do sleep 2; done 可能报 parse error near 'done'(2026-09-02 在一台 macOS zsh 环境实测),用 heredoc 最稳。
  • 亮底插画进深色主题要压暗:filter: brightness(.84) saturate(.92)。
  • 页面接线后 curl -I 每张图 200;仓库结构迁移后 favicon/og 静默 404 是踩过的坑。

与 rankup 的关系

  • rankup/SKILL.md 段 3(建站)硬规则和「一句话落到哪一段」表都指到 /imagegen:网站需要任何视觉素材就来这里真实生成。
  • 段 3 红线:任何页面不得出现占位图(灰块、placeholder、模板示例图),Google 据此判垃圾站、整站连坐;段 4 硬规则:每页独立 og:image 且必须有图,≥1200px 宽,全站共用一张不通过。两条叠加,没有生成能力就只剩占位一条路——所以出图是建站流程的固定环节,不是可选美化。
  • 图做完回 rankup 走闸门:checklists.md 段 3「无占位红线」、段 4「每页独立 title/description/og:image 且有图」、闸门 0 图标全集与 manifest.json 引用全真实、标记经 16px 实测。
  • og 双格式:og:image / twitter:image 用 PNG/JPEG,页内 <img>、JSON-LD image、image-sitemap 用 WebP。

反模式

  • 翻 codex --help 找不到 image flag 就宣布"Codex 不能生图"。
  • 前台跑 codex exec 让用户干等;或紧密轮询后台 shell。
  • 有常设授权的机器上还为 --sandbox danger-full-access 反复请示——按该机器的规则处理一次,启动行披露即可。
  • 提示词里让模型画文字、标语、品牌名。
  • 没用 Read 打开过就把图接进页面;或把未压缩的多 MB 原图提交进仓库。
  • 一页一图全站共用同一张 og:image;或先塞一张灰块"以后再换"。
  • 用单张他人作品做参考并在提示词里点名角色。
  • 只做一两道验收就往下走;非盲判断直接写进结论。
  • 为"省配额"缩小批量或不敢重生成——先看当前账号的计划;额度充足时第一版差一点就再跑,不要拿猜测的配额限制自己。
  • 在 subagent 内部用 run_in_background: true 跑 codex exec——subagent park 后 harness 判定其空闲并终止,完成通知永远送不到,图生成了但没人收。必须从主线程跑 codex exec(Bash 的 run_in_background: true),主线程能正确接收完成通知;或者 subagent 内改用前台同步等待(但会占用 subagent 上下文数分钟)。2026-09-11 同一批任务因此重试三次。

已验证(2026-09-02,codex-cli 0.149.0)

  • 链路:两张图一套(1024² 透明吉祥物 + 1200×630 真人场景 og:image),命令即上文模板,-o final.md 拿最终报告。exit 0,无报错,总耗时 302 s(约 5 分钟)。
  • Codex 用的方法:内置 image_gen 工具(日志里出现 gpt-image-1.5 / gpt-image-2 字样,Codex 自报「具体模型名未公开」);本轮共调了 9 次,原图全部落在 ~/.codex/generated_images/<session-id>/exec-*.png,每张 0.7–1.5 MB。最终文件是它用 ImageMagick(magick,本机 7.1.2)后处理得到的:吉祥物 resize → -remap 锁定提示词里的四个 hex → alpha 提取;og 图轻微居中裁切缩到精确 1200×630。要没被后处理过的原图去 generated_images/ 拿。
  • 结果:og-image.png 1200×630 RGB 757 KB,构图、光线、屏幕模糊无字全部按要求;mascot-logo.png 1024×1024 RGBA 33 KB(四色量化),真透明——57% 像素 alpha=0,alpha 极值 (0,255)。
  • 透明背景有一段弯路:跑到约 3.5 分钟时目录里已有一版 mascot-logo.png,alpha 全 255、深蓝底被烤进去了;Codex 自己又跑了两轮才在第 5 分钟做出真透明。教训两条:① final.md / 退出码出来之前别量图,中间态会误判;② 提示词里 true alpha, not a white/dark square 这句要写,它确实靠这句自检。若最终仍不透明,按「压缩与落盘」的 ImageMagick 抠底命令二次处理,不要写成功。
  • 没达标的一项:吉祥物 alpha 包围盒占比 0.62,提示词写了 75–80% 仍不够(与 2026-08-22 那批 41%–66% 同一现象)。修法在「压缩与落盘」:-trim 后按目标占比 -resize 再 -extent 回 1024²。
  • 压缩实测:og PNG 757 KB → sips JPEG q82 92 KB → cwebp q82 27 KB;四色量化的透明 PNG 33 KB → WebP 反而 48 KB,已量化的扁平透明 PNG 直接用,不转 WebP。
  • 接触印相 sips -Z 120 两张均可辨主体;盲评本次未做(单张无对照)。

© yan-labs, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files in imagegen of yan-labs/yan-skills.

  • SKILL.md
  • README.md
  • evals/evals.json

Open the folder on GitHubat commit 8ec2915

Compare with similar skills

Imagegen next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Imagegen compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Imagegen this skillyan-labs/yan-skills208—~2.5kAutomated safety check: PassMIT
Structured Image Generationbytedance/deer-flow83k5 repos~2.9kAutomated safety check: PassMIT
AI Image Generation and Editingzhayujie/CowAgent47k—~1.3kAutomated safety check: PassMIT
Canghe Comicfreestylefly/canghe-skills4618 repos~3.2kAutomated safety check: PassNone
Generate Imageynulihao/AgentSkillOS61710 repos~1.7kAutomated safety check: NotesNone
GPT Image Generation CLIwuyoscar/GPT-Image2-Skill5.7k—~2.5kAutomated safety check: NotesMIT

Similar skills

  • Structured Image Generation

    bytedance/deer-flow

    Turns an image request into a structured JSON prompt and runs a bundled Python script to generate the picture, optionally guided by reference images.

    83k GitHub starsUsed in 5 repos~2.9k tokens
    Media & CreativeAuto-check passed
  • Generates or edits images from text prompts through a Python script that picks an image backend based on which API keys are configured.

    47k GitHub stars~1.3k tokensUpdated today
    Media & CreativeAuto-check passed
  • Canghe Comic

    freestylefly/canghe-skills

    Knowledge comic creator supporting multiple art styles and tones.

    461 GitHub starsUsed in 8 repos~3.2k tokens
    Media & CreativeAuto-check passed
  • Generate Image

    ynulihao/AgentSkillOS

    Generate or edit images using AI models (FLUX, Gemini). An agent skill from ynulihao/AgentSkillOS.

    617 GitHub starsUsed in 10 repos~1.7k tokens
    Media & CreativeAuto-check: notes
  • GPT Image Generation CLI

    wuyoscar/GPT-Image2-Skill

    Generates and edits images with GPT Image 2 or 2.5 through a packaged CLI and a prompt gallery, after settling which model fits the request.

    5.7k GitHub stars~2.5k tokensUpdated 8 days ago
    Media & CreativeAuto-check: notes
  • Minimal Zine Poster Generator

    LiamGvchi/gc-minimal-zine-poster

    Creates or analyzes quiet, paper-texture zine posters with big negative space, one color accent and experimental type, returning an image prompt and the generated poster.

    7.3k GitHub stars~2.9k tokensUpdated 1 mo ago
    Media & CreativeAuto-check passed

More from yan-labs/yan-skills

  • Rankup

    yan-labs/yan-skills

    独立开发者的项目全生命周期管理:需求验证、选词选品、建站或做原生 App、上线接入、SEO/GEO 获客、支付变现、监控迭代与跨会话接力。以下情况使用:用户提到 rankup 或 /rankup(check、review、init;doctor 仅指整理项目 .rankup 记录);当前目录或工作区有…

    208 GitHub stars~5.1k tokensUpdated 4 days ago
    Auto-check: notes
  • Backlink

    yan-labs/yan-skills

    A skill your agent uses for backlink work - finding, qualifying, submitting, analyzing or verifying backlinks, directory and blog-comment placements, competitor link sources, anchors, toxic links…

    208 GitHub stars~8.9k tokensUpdated 4 days ago
    Auto-check passed
  • Skill Link Check

    yan-labs/yan-skills

    A skill your agent uses when a skill or slash command is missing, duplicated, not loading, or inconsistent between .agents/skills and .claude/skills; includes “skill 没生效”“检查 skill 链接” and broken…

    208 GitHub stars~573 tokensUpdated 4 days ago
    Auto-check passed
  • Codex

    yan-labs/yan-skills

    当用户明确要求直接操作 Codex CLI 作为后台代理(自选 sandbox、codex review、apply、resume 等原生命令,或用 git worktree 并行布置多个 worker)做代码分析、编辑、审查时使用。普通的“让 Codex 或 GPT-6 写代码、调研、review”派单走 agent-fleet(fleet code);图片生成与网站视觉素材走…

    208 GitHub stars~2.2k tokensUpdated 4 days ago
    Auto-check passed
  • Cf CLI

    yan-labs/yan-skills

    A skill your agent uses when the user asks for Cloudflare's cf CLI, cf cli search, or cloudflare.config.ts to inspect resources, manage zones or DNS, attach domains, or deploy.

    208 GitHub stars~1.1k tokensUpdated 4 days ago
    Auto-check passed
  • Autopilot

    yan-labs/yan-skills

    当用户说“autopilot”“自己搞定”“直接跑到底”,或委托代理调查、实施并验证一个宽泛任务时使用。仅规划、解释或已有明确窄范围的请求,按用户指定范围处理。

    208 GitHub stars~616 tokensUpdated 4 days ago
    Auto-check passed

Questions about Imagegen

What does Imagegen do?

当用户要求生成图片、配图、插图、logo、吉祥物、封面、海报、og 图、favicon 源图、真人场景图,或建站时页面缺图、仍有占位图时使用。包括“画一张”“出一套图”“网站需要配图”“image gen”。本 Skill 负责提示词、生成、三道验收、压缩与落盘;只是让 Codex CLI 做代码代理任务时用 codex,普通派单走 agent-fleet,页面 SEO 与上线闸门走…. Imagegen is an agent skill from yan-labs/yan-skills.

When should I use Imagegen?

Imagegen fits situations like: tasks that involve Image generation.

How do I install Imagegen in Claude Code?

Run `npx skills add yan-labs/yan-skills --skill imagegen -a claude-code`. Or copy the skill folder (imagegen in yan-labs/yan-skills) into .claude/skills/imagegen in your project. Claude Code loads it when a task matches its description.

How do I install Imagegen in Codex?

Run `npx skills add yan-labs/yan-skills --skill imagegen -a codex`. Or copy the skill folder (imagegen in yan-labs/yan-skills) into .agents/skills/imagegen in your project. Codex loads it when a task matches its description.

Can I use Imagegen in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add yan-labs/yan-skills --skill imagegen -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/imagegen, .gemini/skills/imagegen, .github/skills/imagegen and .opencode/skills/imagegen in your project.

What does Imagegen need to run?

Going by SKILL.md and its folder, Imagegen needs the command-line tools its instructions call (codex, magick, brew and curl). Our summary lists: Python 3.

Does Imagegen access the network?

SKILL.md contains no URLs. Its commands use curl, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Imagegen safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Imagegen use?

Imagegen is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Imagegen use?

About 2.5k tokens (SKILL.md is roughly 9.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Imagegen?

Skills that share tags, products or a category with Imagegen: Structured Image Generation (bytedance/deer-flow, 83k stars), AI Image Generation and Editing (zhayujie/CowAgent, 47k stars), Canghe Comic (freestylefly/canghe-skills, 461 stars) and Generate Image (ynulihao/AgentSkillOS, 617 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Imagegen?

yan-labs (a GitHub user) maintains it in yan-labs/yan-skills, which has 208 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on October 4, 2026.

Source: yan-labs/yan-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.