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 anymouschina/TapCanvas --skill ecom-details-imagegit clone --depth 1 https://github.com/anymouschina/TapCanvasWrote 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/anymouschina/tapcanvas/ecom-details-image)<a href="https://agentmods.dev/skills/anymouschina/tapcanvas/ecom-details-image"><img src="https://agentmods.dev/badge/skills/anymouschina/tapcanvas/ecom-details-image/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/anymouschina/tapcanvas/ecom-details-image"><img src="https://agentmods.dev/badge/skills/anymouschina/tapcanvas/ecom-details-image.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 | $0.00200 | $0.04646 |
| Opus 5.5 | $0.00080 | $0.01858 |
| Sonnet 5.5 | $0.00040 | $0.00929 |
| Haiku 4.5 | $0.00020 | $0.00465 |
Grade A, and why
ecom-details-image 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 1mo 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 — 287 lines — stays where its author put it; the contents beside it link to each section on GitHub.
电商商品与详情页视觉
这个 Skill 解决的是转化问题,不是审美问题。商品图的成败由「买家 3 秒内是否看懂这是什么、为什么值得买」决定。
两种工作模式:
- Brief / Prompt 模式:只交付视觉简报与可执行图片 prompt。用户没有明确要求出图时默认走这里。
- Generate 模式:用户明确要求“生图/出图/生成图片/render”时,先完成 prompt,再沿 TapCanvas 自有生图链路真实产出图片资产。
边界与交接
- 本 Skill 负责商品、店铺与营销静图的策略、prompt 与图片包编排。
- 视频成片、广告片与带货视频交给
tvc-ad-director/tapcanvas-video-workflow;本 Skill 只提供其中的静帧与 packshot 图方向。 - 小红书/公众号封面(内容型封面,主体是标题与信息密度)交给
xhs-cover-anchor;本 Skill 处理的是商品与店铺视觉。 - 需要设计角色、场景或道具的可复用身份锚时,交给
tapcanvas-character-card/tapcanvas-scene-card/tapcanvas-prop-card,本 Skill 不新建身份卡体系。 - 只问模型参数、渠道限制或接口能力时,读当前动态 schema,不用本 Skill 代答。
不要索取、写入或回显任何 API key。本项目的生图能力由自有链路提供,不需要用户配置外部密钥。
单一操作回环
- 判断图片用途(主图 / 副图 / 详情页屏 / 广告位 / 社媒 / 店铺)与转化驱动力(视觉驱动 / 痛点驱动 / 情感价值驱动,判据见
references/conversion-sequences.md)。 - 按语义从下方场景方向索引选择本轮真正需要的方向,只读取对应的模板文件。
- 只收集会实质改变画面结果的缺失信息;非关键字段明确假设后继续,不无谓阻塞。
- 形成视觉简报:主体、用途、受众语境、风格、构图与比例、图片内文字、否定约束。
- 任务包含多张图时,先建立 Campaign Style Lock,锁定整套图的色板、冷暖调、字体、背景、光线、布局与图标风格。
- 写出可执行 prompt,逐条对照下方铁律;多图任务把同一段 Style Lock 原样放进每张 prompt 的第一段。
- 商品或营销任务先完成转化驱动力诊断,再决定画面顺序,不要从模板堆里挑图型凑数。
- 用户要求整套商品图时,交付 5 张主图 + 7-9 张详情页的图片包计划。
- 用户明确要求出图时按下方执行链路真实提交;否则交付 prompt 包。
出图执行(TapCanvas 唯一链路)
禁止:调用第三方图像 API、注册外部图床、设置任何 *_API_KEY、安装上游脚本、把 base64/临时路径/第三方临时链接当作交付资产。
agents bridge / 小T 会话
tapcanvas_get_tool_schema({name:"tapcanvas_image_generate_to_canvas"})读取当前请求作用域下的动态 schema。tapcanvas_call_tool({name:"tapcanvas_image_generate_to_canvas", args:{...}})逐张提交。
终端 / CLI
统一走 apps/agents-cli/skills/tapcanvas-api(toolExecute → 同一业务工具,或 draw endpoint)。不得绕过该 skill 直接拼接请求。
产品一致性
- 用户提供产品照片时,参考图比文字描述更有效,必须把参考图传入:普通生图用
referenceImageNodeIds/referenceAssetIds,需要区分职责时用referenceAssetBindings:[{assetId,role,strength?}],role取content(产品外观)或identity。 - 只传节点/资产 ID;禁止向 agent 面工具传 URL,也禁止传
referenceImages、styleImages、styleReferenceImages、imageUrl、assetInputs[].url。 - 项目已有全局画风锚时,由服务端在付费边界注入,不要自行复制、丢弃或另造第二套画风。
What ships with it
28 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.
- PROVENANCE.md 2.5 KB
- references/conversion-sequences.md 9.5 KB
- references/prompt-craft.md 3.6 KB
- references/templates/01-hero-image.json 2.4 KB
- references/templates/02-lifestyle-scene.json 2.8 KB
- references/templates/03-flat-lay.json 2.7 KB
- references/templates/04-detail-macro.json 2.5 KB
- references/templates/05-poster-banner.json 2.8 KB
- references/templates/06-social-media.json 3.0 KB
- references/templates/07-ugc-style.json 3.9 KB
- references/templates/08-model-showcase.json 3.1 KB
- references/templates/09-before-after.json 2.3 KB
- references/templates/10-packaging.json 2.4 KB
- references/templates/11-infographic.json 2.5 KB
- references/templates/12-creative-concept.json 2.4 KB
- references/templates/13-size-spec.json 2.4 KB
- references/templates/14-multi-product.json 2.4 KB
- references/templates/15-livestream.json 3.2 KB
- references/templates/16-try-on-virtual.json 2.6 KB
- references/templates/17-exploded-view.json 2.7 KB
- references/templates/18-ghost-mannequin.json 2.8 KB
- references/templates/19-multi-angle-grid.json 2.5 KB
- references/templates/20-magazine-editorial.json 3.0 KB
- references/templates/21-seasonal-campaign.json 2.6 KB
- references/templates/22-luxury-atmospherics.json 2.7 KB
- references/templates/23-device-mockup.json 2.7 KB
- references/templates/24-storefront.json 2.7 KB
- references/templates/25-sports-campaign.json 2.7 KB
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.
- 1mo ago First seen · 287 lines · 200 tokens per session scan A 457c4cd8c5bf
ecom-details-image is a skill published in the GitHub repository anymouschina/TapCanvas (681 stars, last pushed today), licensed MIT. It adds 200 tokens to every session and 4,646 once invoked, about $0.0008 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 2026-09-21.
Other skills, from other repositories
workflows
Use in Freezone/虾画 chat when the user selects a Skill or asks to create, plan, analyze, or expand a dynamic canvas workflow for short drama, ads, ecommerce, product video, MV, text, image, video, or audio production.
ai-model-expert-1688-ecommerce-image-generation-editing
An AI image-generation and editing workflow for 1688 e-commerce content, using text or reference images to guide the result. 1688 is a Chinese online wholesale marketplace.
ai-model-expert-1688-ecommerce-video-generation-editing
An AI tool workflow for creating and editing 1688 e-commerce videos from text, images, videos, or audio. 1688 is a Chinese online wholesale marketplace.
302-ai-alternative-ai-hive
A migration guide for teams moving between AI model API platforms, especially for image and video generation. It helps review an existing OpenAI-compatible integration and plan testing, routing, and fallback options.
ai-hive-animation-167-e6ff654
An AI-HIVE workflow for creating original 8-bit pixel-art animation or game-oriented content. It covers characters, settings, storyboards, key frames, animation clips, and checks for visual consistency and originality.
ai-hive-agentic-ocr-targeted-evidence
A document-search workflow that first finds likely pages and then uses OCR, or software that reads text from images, on only the relevant scans, tables, or small text.