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 ZSeven-W/craft-skills --skill qwen-image-gengit clone --depth 1 https://github.com/ZSeven-W/craft-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/zseven-w/craft-skills/qwen-image-gen)<a href="https://agentmods.dev/skills/zseven-w/craft-skills/qwen-image-gen"><img src="https://agentmods.dev/badge/skills/zseven-w/craft-skills/qwen-image-gen/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/zseven-w/craft-skills/qwen-image-gen"><img src="https://agentmods.dev/badge/skills/zseven-w/craft-skills/qwen-image-gen.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 | $0.00075 | $0.01106 |
| Opus 5.5 | $0.00030 | $0.00442 |
| Sonnet 5.5 | $0.00015 | $0.00221 |
| Haiku 4.5 | $0.00007 | $0.00111 |
Grade A, and why
qwen-image-gen 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 23d 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.
What it actually says
Qwen Image Gen
将创作需求转成可执行的提示词、参考图关系和尺寸决策。默认由当前 Agent 改写，无需额外部署提示词增强模型。参考图任务需要视觉理解能力；纯文本执行者应先取得准确的图片描述。
先区分任务
- 提示词、改写、翻译：交付可复制文本与尺寸建议，不提交生图。
- 文生图：描述最终可见画面。身份、数量、文字、动作、颜色和位置是固定要求；只补全必要的构图与光照细节。
- 修改这张图：先说改哪里、改成什么，再简洁锁定未修改内容。改变姿势时不能同时要求姿势不变。
- 参考主体创作新图：先实际看图，明确哪张图提供身份、服装或背景，再描述新姿势与场景。不要只靠重复人物形容词维持同一张脸。
写提示词前读 rewriting.md。保留用户的创作意图，不用通用美化词覆盖具体要求。简单需求不必扩成长文。
把单张、系列和画幅拆开
单张只安排一个明确动作。系列总提示词包含多个动作时，拆为各自完整的单图提示词；同人物系列分别从一张通过验收的身份参考出发，避免逐张把上次生成的误差累积进去。只有用户要拼图时才描述多面板。
画幅是独立参数。明确像素或比例优先；没有指定时，根据构图建议比例并落实到下游接口。不要只在文字里说竖图，却提交方形画布。相同比例可以对应不同像素尺寸。
可供程序衔接的结构：
{"rewritten_prompt":"最终画面描述","wh_ratio":"9:16"}
{"rewritten_prompt":"具体编辑要求及保留条款","wh_ratio":"","ratio_follow":"<image1>"}
wh_ratio 与 ratio_follow 互斥。仅将 rewritten_prompt 送入模型的文字输入，不把 JSON 包装或解释一并送进去。用户只要提示词时，用普通文本交付即可。
已要求生成时
使用用户指定且已有的 Qwen 入口，沿用当前授权。先读 integration.md，核对实际接口：是否支持参考图、多图、显式宽高和原生 Alpha。模型有某能力，不代表当前封装已经接入。
本 Skill 不预置服务器地址或密钥，也不默认下载模型。没有生成连接时仍可交付提示词；不能把“已准备请求”说成“已生成”。
应用需要加载同一份版本化规则时，读 shared-runtime.md。可选 Node.js 模块只构造规划消息、校验结构和整理质量反馈，不联网、不执行生成；普通 Agent 无需运行它。
附带的 scripts/prepare_workbench_request.py 只是一个可选、离线、特定工作台协议的请求构建器。它不调用模型、不进行提示词改写、不联网。其 2048 像素和单参考图限制仅适用于该适配器，不是 Qwen 的模型上限。详见 workbench-adapter.md。
验收与修正
检查未经修饰的原图：目标文字、动作、肢体连接、身份、保留区域、实际像素尺寸；需要透明时检查 Alpha 和边缘，而非仅看文件扩展名。
为每个结果保留原提示词、实际提交文本、参考图、生成设置及结果位置。质量缺陷与审美不喜欢分开记录：多肢不代表不喜欢坐姿。修正时附上缺陷、目标状态和需要保留的内容；保留原图与新版本，让用户比较验收。没有更多生成授权时停止，不无限重抽。
当用户要 A/B 对照，固定图像模型、尺寸、步数、种子和参考图，只改变待测因素。把采用新画幅的实验单独比较。生成成功、图片更美观、严格完成要求是三个不同判断。
遇到文字泄漏、多宫格、姿势粘连、尺寸异常或透明边缘问题时读 tested-behavior.md。这些是具体条件下的观察，不能扩展为通用成功率。
What ships with it
36 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 325 B
- ASSET-LICENSE.md 1.4 KB
- assets/examples/ab/baker-A.png 1385 KB
- assets/examples/ab/baker-B.png 1331 KB
- assets/examples/ab/baker-reference.png 5012 KB
- assets/examples/ab/cases.json 13 KB
- assets/examples/ab/checksums.json 2.1 KB
- assets/examples/ab/poster-A.png 4062 KB
- assets/examples/ab/poster-B.png 3199 KB
- assets/examples/ab/reach-A.png 1426 KB
- assets/examples/ab/reach-B.png 1495 KB
- assets/examples/ab/README.md 1.5 KB
- assets/examples/ab/stool-A.png 1606 KB
- assets/examples/ab/stool-B.png 1558 KB
- assets/examples/portraits/manifest.json 3.5 KB
- assets/examples/portraits/noir-portrait.png 2910 KB
- assets/examples/portraits/noir-portrait.txt 704 B
- assets/examples/portraits/README.md 612 B
- assets/examples/portraits/rooftop-green.png 2917 KB
- assets/examples/portraits/rooftop-green.txt 755 B
- assets/examples/portraits/strawberry-cake.png 2925 KB
- assets/examples/portraits/strawberry-cake.txt 753 B
- assets/manifest.json 4.5 KB
- LICENSE 1.0 KB
- README.en.md 5.6 KB
- README.md 5.8 KB
- references/integration.md 1.9 KB
- references/rewriting.md 6.5 KB
- references/shared-runtime.md 5.6 KB
- references/tested-behavior.md 3.4 KB
- references/workbench-adapter.md 1.7 KB
- runtime/core.mjs 11 KB runs code
- runtime/manifest.json 185 B
- scripts/prepare_workbench_request.py 6.7 KB runs code
- tests/test_prepare_request.py 5.8 KB runs code
- tests/test_runtime.mjs 9.0 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.
- 23d ago First seen · 56 lines · 75 tokens per session scan A 1d5d16ecef84
qwen-image-gen is a skill published in the GitHub repository ZSeven-W/craft-skills (227 stars, last pushed 23d ago), licensed Apache-2.0. It adds 75 tokens to every session and 1,106 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-23.
Other skills, from other repositories
research-report
Research report on companies, competitors, markets or products: research a company, compare competitors, market landscape, funding, pricing, launches. Also to save a brief, summary or snapshot built from search results.
bundled-keep
Existing task skill that should remain after job-level skill injection.
da-anomaly-analysis
Metric anomaly investigation: DAU, GMV, conversion, retention, revenue, KPI or dashboard metric drop, dropped, fell, spike or spiked. Metric attribution and funnel breakdown.
meeting-prep
Meeting prep: brief me before a meeting, call, standup, 1:1 or pitch — who I'm meeting, background, talking points; prep me, ready.
schedule-meeting
Schedule, book or set up a meeting, call, coffee, lunch, dinner, event or appointment on my calendar.
building-agents
Use when building or restructuring an LLM agent — provider adapter, tool calling, structured output, RAG, agent loop, eval gate, cost routing, tracing, MCP server — model-agnostic across OpenAI/Anthropic/Gemini/OSS so a model swap is a config change. NOT vector-store SQL alone (that is postgresdb) or service…