Spot it. Save it. Free, local-first Chrome extension for images, videos and prompts. Explore 2,000+ curated cases and connect your library to AI Agents via MCP. 提示词导演
JavaScript ai-artai-videochrome-extensionclaude-codecodexhiggsfieldjimenglocal-firstmcpmidjourneymoodboardprompt-engineeringprompt-libraryprompt-managerreference-libraryseedancevisual-bookmarksweb-clipper
import test from "node:test";
import assert from "node:assert/strict";
import { collectPageCaptureSnapshot, PAGE_CAPTURE_ADAPTERS } from "../extension/page-capture.js";
import { resolveCapturedVideoOriginal } from "../extension/captured-video-original.js";
import {
collectPageCaptureSitePayload,
installPageCaptureSiteObserver,
isTrustedPageCaptureMediaUrl,
normalizePageCaptureSitePayload
} from "../extension/page-capture-site-adapters.js";
test("Jimeng structured payload restores author, prompt, model, time, metrics and original media", () => {
const normalized = normalizePageCaptureSitePayload({
adapter: "jimeng",
workId: "7490123456789012345",
modelNames: { jimeng_v40: "即梦 4.0" },
item: {
common_attr: { id: "7490123456789012345", title: "孟菲斯光线4", create_time: 1786000000 },
author: { name: "金戈 米良° · A | © ✨", uid: "creator-7" },
aigc_image_params: { text2image_params: { prompt: "平面影像,低饱和度,动态模糊的光线。" } },
aigc_draft: { content: JSON.stringify({ component_list: [{ abilities: { generate: { core_param: { model: "jimeng_v40" } } } }] }) },
statistic: { favorite_num: 103, usage_num: 12 },
image: { large_images: [{ image_url: "https://p3-dreamina-sign.byteimg.com/original.webp?x=1", width: 2160, height: 3840 }] }
}
}, "https://jimeng.jianying.com/ai-tool/work-detail/7490123456789012345");
assert.equal(normalized.contentText, "平面影像,低饱和度,动态模糊的光线。");
assert.equal(normalized.sourceFacts.author, "金戈 米良° · A | © ✨");
assert.equal(normalized.sourceFacts.model, "即梦 4.0");
assert.equal(normalized.sourceFacts.itemId, "7490123456789012345");
assert.equal(normalized.sourceFacts.dimensions, "2160×3840");
assert.deepEqual(normalized.sourceFacts.engagement, { favorites: 103, uses: 12 });
assert.equal(normalized.media[0].variants[0].sourceKind, "site-original");
assert.equal(normalized.completeness, "complete");
});
test("Jimeng observer keeps a bounded sanitized feed without request credentials", () => {
const injected = (0, eval)(`(${installPageCaptureSiteObserver.toString()})`);
const original = {
location: globalThis.location,
fetch: globalThis.fetch,
XMLHttpRequest: globalThis.XMLHttpRequest,
__get_explore_result: globalThis.__get_explore_result,
__image_generate_model_config__: globalThis.__image_generate_model_config__
};
globalThis.location = { hostname: "jimeng.jianying.com", href: "https://jimeng.jianying.com/ai-tool/home" };
globalThis.fetch = undefined;
globalThis.XMLHttpRequest = undefined;
globalThis.__get_explore_result = { data: { item_list: [
{ common_attr: { id: "7490123456789012345" }, private_token: "must-not-survive", image: { large_images: [] } },
{ common_attr: { id: "7490123456789012346" }, request_headers: { cookie: "must-not-survive" }, image: { large_images: [] } }
] } };
try {
const result = injected({ maxCandidates: 1, maxMedia: 1 });
const state = globalThis.__PROMPTDIRECTOR_JIMENG_CAPTURE__;
assert.equal(result.installed, true);
assert.equal(state.items.length, 1);
assert.equal(state.items[0].common_attr.id, "7490123456789012346");
assert.equal("private_token" in state.items[0], false);
assert.equal("request_headers" in state.items[0], false);
} finally {
delete globalThis.__PROMPTDIRECTOR_JIMENG_CAPTURE__;
Object.assign(globalThis, original);
}
});
test("Jimeng observer accepts a user batch above 100 and retains every sanitized item", () => {
const injected = (0, eval)(`(${installPageCaptureSiteObserver.toString()})`);
const original = {
location: globalThis.location,
fetch: globalThis.fetch,
XMLHttpRequest: globalThis.XMLHttpRequest,
__get_explore_result: globalThis.__get_explore_result,
__image_generate_model_config__: globalThis.__image_generate_model_config__
};
globalThis.location = { hostname: "jimeng.jianying.com", href: "https://jimeng.jianying.com/ai-tool/home" };
globalThis.fetch = undefined;
globalThis.XMLHttpRequest = undefined;
globalThis.__get_explore_result = { data: { item_list: Array.from({ length: 180 }, (_, index) => ({
common_attr: { id: String(7490123456789000000n + BigInt(index)) },
private_token: "must-not-survive", image: { large_images: [] }
})) } };
try {
const result = injected({ maxCandidates: 180, maxMedia: 1 });
const state = globalThis.__PROMPTDIRECTOR_JIMENG_CAPTURE__;
assert.equal(result.installed, true);
assert.equal(state.items.length, 180);
assert.equal(state.items[0].common_attr.id, "7490123456789000000");
assert.equal("private_token" in state.items[0], false);
assert.equal("request_headers" in state.items[0], false);
} finally {
delete globalThis.__PROMPTDIRECTOR_JIMENG_CAPTURE__;
Object.assign(globalThis, original);
}
});
test("Jimeng parser rejects lookalike media hosts and remains explicitly partial when page data is absent", () => {
assert.equal(isTrustedPageCaptureMediaUrl("jimeng", "https://p3-dreamina-sign.byteimg.com/a.webp"), true);
assert.equal(isTrustedPageCaptureMediaUrl("jimeng", "https://p3-dreamina-sign.byteimg.com.evil.example/a.webp"), false);
assert.equal(isTrustedPageCaptureMediaUrl("krea", "https://gen.krea.ai/images/public.png"), true);
assert.equal(isTrustedPageCaptureMediaUrl("krea", "https://gen.krea.ai.evil.example/public.png"), false);
assert.equal(isTrustedPageCaptureMediaUrl("higgsfield", "https://d2ol7oe51mr4n9.cloudfront.net/public.webp"), true);
assert.equal(isTrustedPageCaptureMediaUrl("higgsfield", "https://random.cloudfront.net/public.webp"), false);
assert.equal(isTrustedPageCaptureMediaUrl("pinterest", "https://i.pinimg.com/originals/public.gif"), true);
assert.equal(isTrustedPageCaptureMediaUrl("pinterest", "https://i.pinimg.com.evil.example/public.gif"), false);
const partial = normalizePageCaptureSitePayload({ adapter: "jimeng", workId: "7490123456789012345", status: "partial" }, "https://jimeng.jianying.com/ai-tool/work-detail/7490123456789012345");
assert.equal(partial.completeness, "partial");
assert.equal(partial.sourceFacts.author, undefined);
});
test("the main-world Jimeng reader is self-contained and parses assigned JSON without evaluating scripts", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
globalThis.location = { hostname: "jimeng.jianying.com", pathname: "/ai-tool/work-detail/7490123456789012345" };
globalThis.document = { scripts: [{ textContent: `window.__get_explore_result=${JSON.stringify({ data: { item_list: [{ common_attr: { id: "7490123456789012345" } }] } })};` }] };
try {
const result = injected();
assert.equal(result.item.common_attr.id, "7490123456789012345");
assert.doesNotMatch(collectPageCaptureSitePayload.toString(), /eval\(|new Function/u);
} finally {
Object.assign(globalThis, original);
}
});
test("Jimeng home is a feed, never a fake work named home", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
globalThis.location = { hostname: "jimeng.jianying.com", pathname: "/ai-tool/home" };
globalThis.document = {
scripts: [{ textContent: `window.__get_explore_result=${JSON.stringify({ data: { item_list: [
{ common_attr: { id: "7490123456789012345" } },
{ common_attr: { id: "7490123456789012346" } }
] } })};` }]
};
try {
const result = injected();
assert.equal(result.pageKind, "feed");
assert.equal(result.workId, "");
assert.deepEqual(result.items.map((item) => item.common_attr.id), ["7490123456789012345", "7490123456789012346"]);
} finally {
Object.assign(globalThis, original);
}
});
test("Jimeng feed normalization emits one real candidate per work", () => {
const item = (id, author) => ({
extra: { template_type: "image" },
common_attr: { id, create_time: 1786000000 },
author: { name: author, uid: `${id}-author` },
aigc_image_params: { text2image_params: { prompt: `${author}的提示词` } },
image: { large_images: [{ image_url: `https://p3-dreamina-sign.byteimg.com/${id}.webp`, width: 2048, height: 2048 }] }
});
const normalized = normalizePageCaptureSitePayload({
adapter: "jimeng",
pageKind: "feed",
items: [item("7490123456789012345", "作者甲"), item("7490123456789012346", "作者乙")]
}, "https://jimeng.jianying.com/ai-tool/home");
assert.equal(normalized.pageKind, "feed");
assert.equal(normalized.candidates.length, 2);
assert.equal(normalized.candidates[0].sourceFacts.itemId, "7490123456789012345");
assert.equal(normalized.candidates[0].sourceFacts.author, "作者甲");
assert.equal(normalized.candidates[0].media[0].placement, "inline");
assert.notEqual(normalized.candidates[0].title, "即梦AI - 一站式AI创作平台");
});
test("direct Jimeng detail falls back to the visible work panel without inventing site identity", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
globalThis.location = { hostname: "jimeng.jianying.com", pathname: "/ai-tool/work-detail/7490123456789012345" };
globalThis.document = {
scripts: [],
body: {
innerText: [
"AIGC大叔",
"+ 关注",
"87",
"2026-05-30 | 内容由 AI 生成",
"图片提示词",
"敦煌文化主题场景,全景,暗黑风。",
"图片 4.7 | 9:16 | 更多",
"做同款",
"用作参考图"
].join("\n")
},
querySelectorAll: () => []
};
try {
const payload = injected({ maxCandidates: 100, maxMedia: 24, maxTextCharacters: 16 * 1024 * 1024 });
const normalized = normalizePageCaptureSitePayload(payload, "https://jimeng.jianying.com/ai-tool/work-detail/7490123456789012345");
assert.equal(normalized.sourceFacts.author, "AIGC大叔");
assert.equal(normalized.sourceFacts.model, "4.7");
assert.equal(normalized.sourceFacts.publishedAt.slice(0, 10), "2026-05-30");
assert.equal(normalized.sourceFacts.engagement.favorites, 87);
assert.equal(normalized.contentText, "敦煌文化主题场景,全景,暗黑风。");
assert.equal(normalized.title, "AIGC大叔 · 2026-05-30");
} finally {
Object.assign(globalThis, original);
}
});
test("LiblibAI public work state keeps creator, prompts, model stack, metrics and original media", () => {
const pageUrl = "https://www.liblib.art/imageinfo/f6c920410dd74d37b6680a1ad90dfd53";
const mediaUrl = "https://liblibai-online.liblib.cloud/img/public/original.png";
const normalized = normalizePageCaptureSitePayload({
adapter: "liblibai",
canonicalUrl: pageUrl,
data: {
uuid: "f6c920410dd74d37b6680a1ad90dfd53",
title: "公开作品标题",
createTime: "2024-10-23T08:47:54.000+00:00",
counter: { likeCount: 23, commentCount: 4, hitCount: 208 },
images: [{
id: 5205767,
originalImageUrl: mediaUrl,
imageUrl: "https://images-wm.liblib.cloud/img/public/watermark.png",
width: 1200,
height: 2048,
generateInfo: { prompt: "公开正向提示词", negativePrompt: "公开负向提示词" },
models: [
{ modelName: "F.1 基础模型", versionName: "v1" },
{ modelName: "细节 LoRA", versionName: "v2" }
]
}]
},
author: { nickname: "脱敏作者", uuid: "public-author-id" }
}, pageUrl);
assert.equal(normalized.adapter, "liblibai");
assert.equal(normalized.contentText, "公开正向提示词");
assert.equal(normalized.sourceFacts.author, "脱敏作者");
assert.equal(normalized.sourceFacts.handle, "public-author-id");
assert.equal(normalized.sourceFacts.model, "F.1 基础模型 v1 · 细节 LoRA v2");
assert.equal(normalized.sourceFacts.dimensions, "1200×2048");
assert.deepEqual(normalized.sourceFacts.engagement, { likes: 23, comments: 4, views: 208 });
assert.equal(normalized.media[0].url, mediaUrl);
assert.equal(normalized.media[0].sourceKind, "site-original");
assert.equal(normalized.completeness, "complete");
});
test("LiblibAI collector reads only bounded public Next data from an image detail", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const pageData = {
props: { pageProps: {
data: {
uuid: "f6c920410dd74d37b6680a1ad90dfd53",
title: "公开作品标题",
createTime: "2024-10-23T08:47:54.000+00:00",
counter: { likeCount: 23 },
images: [{
id: 5205767,
originalImageUrl: "https://liblibai-online.liblib.cloud/img/public/original.png",
imageUrl: "https://images-wm.liblib.cloud/img/public/watermark.png",
width: 1200,
height: 2048,
generateInfo: { prompt: "公开正向提示词", metainformation: "must-not-survive" },
models: [{ modelName: "F.1 基础模型", versionName: "v1", privatePath: "must-not-survive" }]
}]
},
authorInfo: { userDetail: { nickname: "脱敏作者", uuid: "public-author-id", mobile: "must-not-survive" } }
} }
};
globalThis.location = {
hostname: "www.liblib.art",
pathname: "/imageinfo/f6c920410dd74d37b6680a1ad90dfd53",
href: "https://www.liblib.art/imageinfo/f6c920410dd74d37b6680a1ad90dfd53"
};
globalThis.document = {
querySelector: (selector) => selector === "#__NEXT_DATA__" ? { textContent: JSON.stringify(pageData) } : null,
querySelectorAll: () => [],
scripts: []
};
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.adapter, "liblibai");
assert.equal(payload.data.images.length, 1);
assert.equal(payload.data.images[0].generateInfo.prompt, "公开正向提示词");
assert.equal(JSON.stringify(payload).includes("must-not-survive"), false);
} finally {
Object.assign(globalThis, original);
}
});
test("Krea public gallery detail keeps the full prompt, UUID, date, dimensions and original image", () => {
const pageUrl = "https://www.krea.ai/feed/vintage-film-portrait-554bfb91-2a06-5e91-8a21-24a162f3b81f";
const mediaUrl = "https://gen.krea.ai/images/554bfb91-2a06-5e91-8a21-24a162f3b81f.png";
const normalized = normalizePageCaptureSitePayload({
adapter: "krea",
pageKind: "detail",
canonicalUrl: pageUrl,
prompt: "A close-up vintage film portrait with warm sunlight and heavy grain.",
jsonLd: [{
"@type": "ImageObject",
name: "A close-up vintage film portrait...",
url: pageUrl,
contentUrl: mediaUrl,
datePublished: "2026-05-13T06:47:52.874Z",
width: 768,
height: 1376
}]
}, pageUrl);
assert.equal(normalized.adapter, "krea");
assert.equal(normalized.contentText, "A close-up vintage film portrait with warm sunlight and heavy grain.");
assert.equal(normalized.sourceFacts.itemId, "554bfb91-2a06-5e91-8a21-24a162f3b81f");
assert.equal(normalized.sourceFacts.publishedAt, "2026-05-13T06:47:52.874Z");
assert.equal(normalized.sourceFacts.dimensions, "768×1376");
assert.equal(normalized.media[0].url, mediaUrl);
assert.equal(normalized.completeness, "complete");
});
test("Krea gallery feed emits bounded real artwork candidates instead of application-shell text", () => {
const normalized = normalizePageCaptureSitePayload({
adapter: "krea",
pageKind: "feed",
items: [{
itemId: "72324387-04f9-526e-ae96-e132f34ab60c",
canonicalUrl: "https://www.krea.ai/feed/a-vivid-green-praying-mantis-72324387-04f9-526e-ae96-e132f34ab60c",
title: "a vivid green praying mantis",
imageUrl: "https://gen.krea.ai/images/72324387-04f9-526e-ae96-e132f34ab60c.png"
}]
}, "https://www.krea.ai/feed");
assert.equal(normalized.pageKind, "feed");
assert.equal(normalized.candidates.length, 1);
assert.equal(normalized.candidates[0].sourceFacts.itemId, "72324387-04f9-526e-ae96-e132f34ab60c");
assert.equal(normalized.candidates[0].media[0].sourceKind, "site-original");
assert.equal(normalized.candidates[0].completeness, "partial");
});
test("Krea collector reads a public detail prompt and ImageObject without application state", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const pageUrl = "https://www.krea.ai/feed/vintage-film-portrait-554bfb91-2a06-5e91-8a21-24a162f3b81f";
const imageObject = {
"@context": "https://schema.org",
"@type": "ImageObject",
name: "Vintage film portrait",
url: pageUrl,
contentUrl: "https://gen.krea.ai/images/554bfb91-2a06-5e91-8a21-24a162f3b81f.png",
datePublished: "2026-05-13T06:47:52.874Z",
width: 768,
height: 1376,
privateSession: "must-not-survive"
};
globalThis.location = { hostname: "www.krea.ai", pathname: new URL(pageUrl).pathname, href: pageUrl };
globalThis.document = {
querySelector: (selector) => selector === "article h1" ? { textContent: "A complete public image prompt." } : null,
querySelectorAll: (selector) => selector === 'script[type="application/ld+json"]'
? [{ textContent: JSON.stringify(imageObject) }]
: []
};
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.adapter, "krea");
assert.equal(payload.pageKind, "detail");
assert.equal(payload.prompt, "A complete public image prompt.");
assert.equal(payload.jsonLd[0].contentUrl, imageObject.contentUrl);
assert.equal(JSON.stringify(payload).includes("must-not-survive"), false);
} finally {
Object.assign(globalThis, original);
}
});
test("Krea collector converts only observed gallery cards to their verified original URL", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const itemId = "72324387-04f9-526e-ae96-e132f34ab60c";
const href = `/feed/a-vivid-green-praying-mantis-${itemId}`;
const imageNode = {
style: { backgroundImage: `image-set(url("https://optim-images.krea.ai/https---gen-krea-ai-images-${itemId}-png-512.webp") 1x, url("https://optim-images.krea.ai/https---gen-krea-ai-images-${itemId}-png-1024.webp") 2x)` }
};
globalThis.location = { hostname: "www.krea.ai", pathname: "/feed", href: "https://www.krea.ai/feed" };
globalThis.document = {
querySelector: () => null,
querySelectorAll: (selector) => selector.startsWith('script[type=')
? []
: [{ href, querySelector: () => imageNode, getAttribute: () => href }]
};
try {
const payload = injected({ maxCandidates: 1, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.items.length, 1);
assert.equal(payload.items[0].itemId, itemId);
assert.equal(payload.items[0].imageUrl, `https://gen.krea.ai/images/${itemId}.png`);
assert.equal(payload.items[0].title, "a vivid green praying mantis");
} finally {
Object.assign(globalThis, original);
}
});
test("Higgsfield project keeps public creator, project brief, date, metrics and artwork", () => {
const pageUrl = "https://higgsfield.ai/@public-creator/projects/anime-production";
const mediaUrl = "https://d2ol7oe51mr4n9.cloudfront.net/public/project-cover.webp";
const normalized = normalizePageCaptureSitePayload({
adapter: "higgsfield",
pageKind: "detail",
canonicalUrl: pageUrl,
brief: "A cinematic anime production brief with deliberate lighting and motion.",
jsonLd: [{
"@type": "CreativeWork",
name: "Anime Production",
description: "Public project description",
url: pageUrl,
author: { "@type": "Person", name: "Public Creator", alternateName: "@public-creator", identifier: "creator-public-id" },
datePublished: "2026-06-18T12:30:00.000Z",
image: mediaUrl,
interactionStatistic: [
{ interactionType: { "@type": "ViewAction" }, userInteractionCount: 812 },
{ interactionType: { "@type": "LikeAction" }, userInteractionCount: 37 }
]
}]
}, pageUrl);
assert.equal(normalized.adapter, "higgsfield");
assert.equal(normalized.title, "Anime Production");
assert.equal(normalized.contentText, "A cinematic anime production brief with deliberate lighting and motion.");
assert.equal(normalized.sourceFacts.author, "Public Creator");
assert.equal(normalized.sourceFacts.handle, "public-creator");
assert.equal(normalized.sourceFacts.itemId, "public-creator/anime-production");
assert.equal(normalized.sourceFacts.publishedAt, "2026-06-18T12:30:00.000Z");
assert.deepEqual(normalized.sourceFacts.engagement, { likes: 37, views: 812 });
assert.equal(normalized.media[0].url, mediaUrl);
assert.equal(normalized.completeness, "complete");
});
test("Higgsfield public community ItemList emits projects without page-shell copy", () => {
const normalized = normalizePageCaptureSitePayload({
adapter: "higgsfield",
pageKind: "feed",
canonicalUrl: "https://higgsfield.ai/community",
jsonLd: [{
"@type": "ItemList",
itemListElement: [{
position: 1,
item: {
"@type": "CreativeWork",
name: "Fashion Motion Study",
description: "A public motion project.",
url: "https://higgsfield.ai/@public-creator/projects/fashion-motion-study",
author: { name: "Public Creator", alternateName: "@public-creator" },
image: "https://d2ol7oe51mr4n9.cloudfront.net/public/fashion-cover.webp"
}
}]
}]
}, "https://higgsfield.ai/community");
assert.equal(normalized.pageKind, "feed");
assert.equal(normalized.candidates.length, 1);
assert.equal(normalized.candidates[0].sourceFacts.author, "Public Creator");
assert.equal(normalized.candidates[0].sourceFacts.itemId, "public-creator/fashion-motion-study");
assert.equal(normalized.candidates[0].title, "Fashion Motion Study");
assert.doesNotMatch(normalized.candidates[0].contentText, /create|sign in|pricing/iu);
});
test("Higgsfield public player records bind original video files to their own playback sources without duplicating JSON-LD media", async () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const snapshot = (0, eval)(`(${collectPageCaptureSnapshot.toString()})`);
const keys = ["location", "document", "window", "chrome", "Readability", "Element", "HTMLCanvasElement"];
const original = Object.fromEntries(keys.map(key => [key, Object.getOwnPropertyDescriptor(globalThis, key)]));
const pageUrl = "https://higgsfield.ai/@public-creator/projects/film";
const poster = "https://d2ol7oe51mr4n9.cloudfront.net/public/shared-poster.png";
// Same public project.media shape as the SSR project and hydrated player.
const project = { media: ["film", "second-film", "unrelated"].map(id => ({
id, kind: "video", src: `https://cdn.higgsfield.ai/hls/video_input/${id}/index.m3u8`,
originalSrc: `https://d2ol7oe51mr4n9.cloudfront.net/public/${id}.mp4`,
thumbnail: poster, width: 3840, height: 2160, accountSession: "PRIVATE_MUST_NOT_LEAK"
})) };
const videos = project.media.slice(0, 2).map(item => ({ "@type": "VideoObject", contentUrl: item.src, thumbnailUrl: poster }));
const work = { "@type": "CreativeWork", url: pageUrl, name: "Film", description: "Complete project description",
author: { name: "Public creator" }, video: videos };
const fiber = { memoizedProps: { project, privateAccount: "PRIVATE_MUST_NOT_LEAK" } };
fiber.return = fiber; // A malformed parent cycle must not hang capture.
const player = { __reactFiber$fixture: { memoizedProps: { src: videos[0].contentUrl }, return: fiber } };
const body = { innerText: work.description, scrollHeight: 800, querySelector: () => null, querySelectorAll: () => [] };
Object.assign(globalThis, { location: new URL(pageUrl), chrome: undefined, Readability: undefined,
window: { scrollX: 0, scrollY: 0, scrollTo: () => {} },
document: { title: work.name, body, baseURI: pageUrl, documentElement: { scrollHeight: 800 },
querySelector: () => null,
querySelectorAll: selector => selector === "video" ? [player]
: selector === 'script[type="application/ld+json"]' ? [{ textContent: JSON.stringify(work) }] : [] }
});
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert(!JSON.stringify(payload).includes("PRIVATE_MUST_NOT_LEAK"));
const siteData = normalizePageCaptureSitePayload(payload, pageUrl);
assert.equal(siteData.media.length, 2, "Only videos declared by this work enter the capture");
assert.deepEqual(siteData.media.map(item => item.url), project.media.slice(0, 2).map(item => item.originalSrc));
assert.deepEqual(siteData.media.map(item => item.streamUrl), project.media.slice(0, 2).map(item => item.src));
assert(siteData.media.every(item => item.width === 3840 && item.height === 2160));
const captured = await snapshot({ adapters: PAGE_CAPTURE_ADAPTERS, siteData });
assert.equal(captured.candidates[0].media.length, 2, "Playback JSON-LD copies must not reappear as extra videos");
assert.deepEqual(captured.candidates[0].media.map(item => item.url), project.media.slice(0, 2).map(item => item.originalSrc));
const requested = [];
const bytes = new Blob(["fixture-original"], { type: "video/mp4" });
const resolved = await resolveCapturedVideoOriginal(captured.candidates[0].media[0], captured.candidates[0], {
download: async url => { requested.push(url); return bytes; }
});
assert.equal(resolved, bytes);
assert.deepEqual(requested, [project.media[0].originalSrc], "The shared save path downloads the original, not the HLS manifest");
// A distinct source with the same cover still remains a distinct video.
const partialSiteData = { ...siteData, media: siteData.media.slice(0, 1) };
const partial = await snapshot({ adapters: PAGE_CAPTURE_ADAPTERS, siteData: partialSiteData });
assert.equal(partial.candidates[0].media.length, 2);
assert.deepEqual(partial.candidates[0].media.map(item => item.url), [project.media[0].originalSrc, project.media[1].src]);
// Exercise real snapshot DOM/structured pairing, with reverse DOM order so
// a shared poster cannot accidentally bind both originals to the first player.
class Element {}
globalThis.Element = Element;
globalThis.HTMLCanvasElement = class extends Element {};
body.childNodes = project.media.slice(0, 2).reverse().map(item => Object.assign(new Element(), {
localName: "video", tagName: "VIDEO", currentSrc: item.src, src: item.src, poster,
videoWidth: 1280, videoHeight: 720, childNodes: [],
matches: selector => selector.split(",").includes("video"), closest: () => null,
getAttribute: name => name === "src" ? item.src : "", querySelectorAll: () => [],
getBoundingClientRect: () => ({ x: 0, y: 0, top: 0, bottom: 720, width: 1280, height: 720 })
}));
const withPlayers = await snapshot({ adapters: PAGE_CAPTURE_ADAPTERS, siteData });
assert.equal(withPlayers.captureError, undefined);
assert.equal(withPlayers.candidates[0].media.length, 2);
for (const [index, media] of withPlayers.candidates[0].media.entries()) {
assert.equal(media.url, project.media[index].originalSrc);
assert(media.variants.every(variant => variant.url === project.media[index].originalSrc || variant.url === project.media[index].src),
"Each original must retain only its own actual playback variants");
}
// Untrusted originals never displace the trusted playback source.
project.media[0].originalSrc = "https://unrelated.example/private.mp4";
const rejected = normalizePageCaptureSitePayload(injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 }), pageUrl);
assert.equal(rejected.media[0].url, project.media[0].src);
assert(!JSON.stringify(rejected).includes("unrelated.example"));
} finally {
for (const key of keys) {
if (original[key]) Object.defineProperty(globalThis, key, original[key]);
else delete globalThis[key];
}
}
});
test("Higgsfield collector exposes only bounded public CreativeWork JSON-LD and the project brief", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const pageUrl = "https://higgsfield.ai/@public-creator/projects/anime-production";
const creativeWork = {
"@context": "https://schema.org",
"@type": "CreativeWork",
name: "Anime Production",
description: "Public project description",
url: pageUrl,
author: { "@type": "Person", name: "Public Creator", alternateName: "@public-creator", privateToken: "must-not-survive" },
datePublished: "2026-06-18T12:30:00.000Z",
image: "https://d2ol7oe51mr4n9.cloudfront.net/public/project-cover.webp",
interactionStatistic: [{ interactionType: { "@type": "ViewAction" }, userInteractionCount: 812 }],
accountSession: "must-not-survive"
};
globalThis.location = { hostname: "higgsfield.ai", pathname: "/@public-creator/projects/anime-production", href: pageUrl };
globalThis.document = {
querySelector: (selector) => selector === '[aria-label^="Project brief:"]'
? { innerText: "A cinematic anime production brief." }
: null,
querySelectorAll: (selector) => selector === 'script[type="application/ld+json"]'
? [{ textContent: JSON.stringify(creativeWork) }]
: []
};
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.adapter, "higgsfield");
assert.equal(payload.pageKind, "detail");
assert.equal(payload.brief, "A cinematic anime production brief.");
assert.equal(payload.jsonLd[0].author.name, "Public Creator");
assert.equal(JSON.stringify(payload).includes("must-not-survive"), false);
} finally {
Object.assign(globalThis, original);
}
});
test("Behance VisualArtwork keeps project owners, description, identity, metrics and cover media", () => {
const pageUrl = "https://www.behance.net/gallery/252610351/Cantor8";
const mediaUrl = "https://mir-s3-cdn-cf.behance.net/projects/404/e4dcec252610351.Y3JvcCwxODQwLDE0MzksMCww.jpg";
const normalized = normalizePageCaptureSitePayload({
adapter: "behance",
pageKind: "detail",
canonicalUrl: pageUrl,
jsonLd: [{
"@type": "VisualArtwork",
name: "Cantor8",
description: "A public digital-design project description.",
url: pageUrl,
identifier: 252610351,
image: mediaUrl,
creator: [
{ "@type": "Person", identifier: 173435329, name: "Public Owner One", url: "https://www.behance.net/publicownerone" },
{ "@type": "Person", identifier: 1443573125, name: "Public Owner Two", url: "https://www.behance.net/publicownertwo" }
],
interactionStatistic: [
{ interactionType: "http://schema.org/LikeAction", userInteractionCount: 766 },
{ interactionType: "http://schema.org/WatchAction", userInteractionCount: 8565 }
]
}]
}, pageUrl);
assert.equal(normalized.adapter, "behance");
assert.equal(normalized.sourceFacts.itemId, "252610351");
assert.equal(normalized.sourceFacts.author, "Public Owner One, Public Owner Two");
assert.equal(normalized.sourceFacts.handle, "publicownerone");
assert.deepEqual(normalized.sourceFacts.engagement, { likes: 766, views: 8565 });
assert.equal(normalized.media[0].url, mediaUrl);
assert.equal(normalized.media[0].sourceKind, "structured");
assert.equal(normalized.completeness, "complete");
});
test("Behance collector reads only the public VisualArtwork schema from a project page", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const pageUrl = "https://www.behance.net/gallery/252610351/Cantor8";
const artwork = {
"@context": "http://schema.org",
"@type": "VisualArtwork",
name: "Cantor8",
description: "A public digital-design project description.",
url: pageUrl,
identifier: 252610351,
image: "https://mir-s3-cdn-cf.behance.net/projects/404/e4dcec252610351.Y3JvcCwxODQwLDE0MzksMCww.jpg",
creator: [{ "@type": "Person", identifier: 173435329, name: "Public Owner", url: "https://www.behance.net/publicowner", email: "must-not-survive" }],
interactionStatistic: [{ "@type": "InteractionCounter", interactionType: "http://schema.org/LikeAction", userInteractionCount: 766 }],
privateSession: "must-not-survive"
};
globalThis.location = { hostname: "www.behance.net", pathname: "/gallery/252610351/Cantor8", href: pageUrl };
globalThis.document = {
querySelector: () => null,
querySelectorAll: (selector) => selector === 'script[type="application/ld+json"]'
? [{ textContent: JSON.stringify(artwork) }]
: []
};
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.adapter, "behance");
assert.equal(payload.jsonLd[0].creator[0].name, "Public Owner");
assert.equal(JSON.stringify(payload).includes("must-not-survive"), false);
} finally {
Object.assign(globalThis, original);
}
});
test("Pinterest Pin keeps the public creator, timestamp, description and original media variants", () => {
const pageUrl = "https://www.pinterest.com/pin/579205202062276414/";
const originalUrl = "https://i.pinimg.com/originals/0d/d6/0c/public-pin.gif";
const normalized = normalizePageCaptureSitePayload({
adapter: "pinterest",
pageKind: "detail",
canonicalUrl: pageUrl,
pin: {
entityId: "579205202062276414",
title: "2026 Web Design Trends to Watch For",
description: "A public design-inspiration Pin description.",
createdAt: "Sat, 21 Dec 2019 21:27:05 +0000",
author: { fullName: "Public Pinner", username: "publicpinner" },
repins: 14,
images: {
original: { url: originalUrl },
large: { url: "https://i.pinimg.com/1200x/0d/d6/0c/public-pin.jpg", width: 1200, height: 592 },
display: { url: "https://i.pinimg.com/736x/0d/d6/0c/public-pin.jpg", width: 600, height: 296 }
}
}
}, pageUrl);
assert.equal(normalized.adapter, "pinterest");
assert.equal(normalized.sourceFacts.itemId, "579205202062276414");
assert.equal(normalized.sourceFacts.author, "Public Pinner");
assert.equal(normalized.sourceFacts.handle, "publicpinner");
assert.equal(normalized.sourceFacts.publishedAt, "2019-12-21T21:27:05.000Z");
assert.deepEqual(normalized.sourceFacts.engagement, { repins: 14 });
assert.equal(normalized.media[0].url, originalUrl);
assert.deepEqual(normalized.media[0].variants.map((variant) => variant.sourceKind), ["site-original", "responsive", "current"]);
assert.equal(normalized.completeness, "complete");
});
test("Pinterest collector sanitizes one public Pin relay payload without retaining request context", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const pageUrl = "https://www.pinterest.com/pin/579205202062276414/";
const response = { data: { v3GetPinQueryv2: { data: {
entityId: "579205202062276414",
gridTitle: "2026 Web Design Trends to Watch For",
gridDescription: "A public design-inspiration Pin description.",
createdAt: "Sat, 21 Dec 2019 21:27:05 +0000",
nativeCreator: { fullName: "Public Pinner", username: "publicpinner", websiteUrl: "must-not-survive" },
images_orig: { url: "https://i.pinimg.com/originals/0d/d6/0c/public-pin.gif" },
images_1200x: { url: "https://i.pinimg.com/1200x/0d/d6/0c/public-pin.jpg", width: 1200, height: 592 },
images_736x: { url: "https://i.pinimg.com/736x/0d/d6/0c/public-pin.jpg", width: 600, height: 296 },
trackingParams: "must-not-survive"
} } } };
globalThis.location = { hostname: "www.pinterest.com", pathname: "/pin/579205202062276414/", href: pageUrl };
globalThis.document = {
scripts: [{ textContent: `window.__PWS_RELAY_REGISTER_COMPLETED_REQUEST__("%7B%22pinId%22%3A%22579205202062276414%22%7D", ${JSON.stringify(response)});` }],
querySelector: (selector) => selector.includes("pinterestapp:repins") ? { content: "14" } : null,
querySelectorAll: () => []
};
try {
const payload = injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 });
assert.equal(payload.adapter, "pinterest");
assert.equal(payload.pin.author.username, "publicpinner");
assert.equal(payload.pin.images.original.url, "https://i.pinimg.com/originals/0d/d6/0c/public-pin.gif");
assert.equal(JSON.stringify(payload).includes("must-not-survive"), false);
} finally {
Object.assign(globalThis, original);
}
});
test("WeChat public article fields retain account, publication time, body and delayed originals", () => {
const pageUrl = "https://mp.weixin.qq.com/s/example";
const normalized = normalizePageCaptureSitePayload({
adapter: "wechat",
canonicalUrl: pageUrl,
title: "视觉案例周刊",
author: "AI Art Works",
publishedAt: "2026-08-01T08:00:00+08:00",
contentText: "第一段公开正文\n\n第二段公开正文",
images: [{
url: "https://mmbiz.qpic.cn/mmbiz_jpg/example/0?wx_fmt=jpeg",
width: 1080,
height: 1440,
alt: "案例作品"
}]
}, pageUrl);
assert.equal(normalized.pageType, "article");
assert.equal(normalized.title, "视觉案例周刊");
assert.equal(normalized.sourceFacts.author, "AI Art Works");
assert.equal(normalized.sourceFacts.publishedAt, "2026-08-01T00:00:00.000Z");
assert.equal(normalized.contentText, "第一段公开正文\n\n第二段公开正文");
assert.equal(normalized.media[0].variants[0].sourceKind, "site-original");
});
test("site collectors expose only bounded public DOM and JSON-LD data outside Jimeng", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
const content = {
innerText: "公开文章正文",
querySelectorAll: () => [{
currentSrc: "https://mmbiz.qpic.cn/mmbiz_jpg/example/0?wx_fmt=jpeg",
src: "",
naturalWidth: 1080,
naturalHeight: 1440,
alt: "案例作品",
getAttribute: (name) => ({ "data-src": "https://mmbiz.qpic.cn/mmbiz_jpg/example/0?wx_fmt=jpeg", "data-w": "1080" })[name] || ""
}]
};
globalThis.location = { hostname: "mp.weixin.qq.com", pathname: "/s/example", href: "https://mp.weixin.qq.com/s/example" };
globalThis.document = {
scripts: [],
querySelector: (selector) => ({
"#js_content": content,
"#activity-name": { textContent: "视觉案例周刊" },
"#js_name": { textContent: "AI Art Works" },
"#publish_time": { textContent: "2026-08-01 08:00" }
})[selector] || null
};
try {
const payload = injected({ maxCandidates: 100, maxMedia: 24, maxTextCharacters: 16 * 1024 * 1024 });
assert.equal(payload.adapter, "wechat");
assert.equal(payload.images.length, 1);
assert.equal(payload.title, "视觉案例周刊");
assert.equal("cookie" in payload, false);
} finally {
Object.assign(globalThis, original);
}
});
test("YesAnd has no dedicated collector and falls through to the generic page engine", () => {
const injected = (0, eval)(`(${collectPageCaptureSitePayload.toString()})`);
const original = { location: globalThis.location, document: globalThis.document };
globalThis.location = { hostname: "yesand.ai", pathname: "/prompt/example", href: "https://yesand.ai/prompt/example" };
globalThis.document = { querySelector: () => null, querySelectorAll: () => [], scripts: [] };
try {
assert.equal(injected({ maxCandidates: 20, maxMedia: 8, maxTextCharacters: 10000 }), null);
} finally {
Object.assign(globalThis, original);
}
});
test("Krea detail keeps original and preview together and does not assign the work prompt to style references", () => {
const url = "https://www.krea.ai/feed/work-554bfb91-2a06-5e91-8a21-24a162f3b81f";
const original = "https://gen.krea.ai/images/554bfb91-2a06-5e91-8a21-24a162f3b81f.png";
const preview = "https://optim-images.krea.ai/preview.webp";
const reference = "https://app-uploads.krea.ai/style.webp";
const result = normalizePageCaptureSitePayload({ adapter: "krea", pageKind: "detail", canonicalUrl: url,
prompt: "A cinematic silhouette.", imagePreviewUrl: preview,
jsonLd: [{ "@type": "ImageObject", contentUrl: original }],
styleReferences: [{ url: reference, previewUrl: "https://www.krea.ai/api/img?fixture", label: "style reference image 1" }]
}, url);
assert.equal(result.pageKind, "detail");
assert.deepEqual(result.media.map(media => media.url), [original, reference]);
assert.deepEqual(result.media[0].variants.map(variant => variant.url), [original, preview]);
assert.equal(result.media[0].originalPrompt, "A cinematic silhouette.");
assert.equal(result.media[1].originalPrompt, undefined);
});
test("untitled Pins use their own description or identity rather than repeating the creator name", () => {
const pin = (id, description) => normalizePageCaptureSitePayload({
adapter:'pinterest', canonicalUrl:`https://www.pinterest.com/pin/${id}/`,
pin:{entityId:id,description,author:{fullName:'Same creator'},images:{original:{url:'https://i.pinimg.com/originals/work.png'}}}
});
assert.equal(pin('123', 'Forest\n temple').title, 'Forest temple');
assert.equal(pin('123', '').title, 'Pinterest 123');
assert.equal(pin('456', '').title, 'Pinterest 456');
assert.equal(pin('456', '').sourceFacts.author, 'Same creator');
});