Token导航 LogoToken导航TokenDH.com
开发需要联网github未标认证来源可访问许可证需确认审计提醒

pixijs-assets像素资产

Agent Skill

pixijs-assets 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

7,704

周安装

321

GitHub Stars

164

下载量

2,568
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:pixijs-assets(像素资产)
来源仓库:https://github.com/pixijs/pixijs-skills
仓库路径:skills/pixijs-assets
安装命令:
npx skills add https://github.com/pixijs/pixijs-skills --skill pixijs-assets
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/pixijs/pixijs-skills --skill pixijs-assets

简介

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合围绕仓库状态、代码变更或协作事项进行整理和分析。
  • 通过 npx skills add 命令安装,需结合原始 README 核验具体用法。
  • 安装前建议确认权限范围和维护状态,避免触发不必要的联网或文件操作。
  • pixijs-assets 属于开发类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

The Assets API is PixiJS's asset loader, resolver, and cache in one singleton. Use it to load textures, video, spritesheets, fonts, JSON, and other resources with format detection, resolution switching, bundle grouping, progress tracking, and GPU cleanup.

Quick Start

await Assets.init({ basePath: "/static/" });

const texture = await Assets.load("bunny.png");
const sprite = new Sprite(texture);
app.stage.addChild(sprite);

const [hero, enemy] = await Assets.load(["hero.png", "enemy.png"]);

await Assets.load({
  alias: "logo",
  src: "logo.webp",
});

const logo = new Sprite(Assets.get("logo"));

Assets.init() is optional but recommended for setting basePath, texturePreference, or a manifest. After init, call Assets.load() with a URL, alias, array, or UnresolvedAsset; resolved assets are cached and re-resolved by Assets.get().

Supported file types

TypeExtensionsParser IDLoader
Textures.png, .jpg, .jpeg, .webp, .aviftextureloadTextures
SVG.svgsvgloadSvg (see references/svg.md)
Video textures.mp4, .m4v, .webm, .ogg, .ogv, .h264, .avi, .movvideoloadVideoTextures (see references/video.md)
Sprite sheets.json (Spritesheet format)spritesheetspritesheetAsset (see references/spritesheet.md)
Bitmap fonts.fnt, .xmlbitmap-fontloadBitmapFont (loading works by default; rendering BitmapText requires 'pixi.js/text-bitmap'; see references/fonts.md)
Web fonts.ttf, .otf, .woff, .woff2web-fontloadWebFont (see references/fonts.md)
JSON.jsonjsonloadJson
Text.txttextloadTxt
Compressed textures.basis, .dds, .ktx, .ktx2basis, dds, ktx, ktx2See references/compressed-textures.md
Animated GIFs.gifgifRequires 'pixi.js/gif'; returns GifSource (see references/gif.md)

The Parser ID column is the value you pass to the top-level parser field on an asset descriptor to force a specific loader. See "Forcing a parser" below.

Forcing a parser with parser

By default, PixiJS picks a loader by matching the file extension or MIME type. When your URL lacks an extension (CDN signed URLs, blob URLs, API endpoints, content-hashed paths), the resolver can't tell the loader what to do. Set the top-level parser field on the asset descriptor to force a specific loader:

// Signed CDN URL with no extension
const texture = await Assets.load({
  src: "https://cdn.example.com/signed/abc123?token=xyz",
  parser: "texture",
});

// API endpoint that returns JSON
const data = await Assets.load({
  alias: "config",
  src: "https://api.example.com/v1/config",
  parser: "json",
});

// Extension-less font URL with explicit family
await Assets.load({
  src: "https://cdn.example.com/fonts/hero-v2",
  parser: "web-font",
  data: { family: "Hero", weights: ["400", "700"] },
});

// Video stream without a file extension
const clipTexture = await Assets.load({
  src: "https://cdn.example.com/stream/xyz",
  parser: "video",
  data: { mime: "video/mp4", muted: true, playsinline: true },
});

The parser field goes at the top level of the asset descriptor (alongside src and data), not inside data. It takes any parser ID from the "Supported file types" table above:

  • 'texture', 'svg', 'video': image, SVG, and video textures
  • 'json', 'text': JSON and plain text
  • 'web-font', 'bitmap-font': web and bitmap fonts
  • 'spritesheet': texture atlas JSON
  • 'gif': animated GIFs (requires 'pixi.js/gif')
  • 'basis', 'dds', 'ktx', 'ktx2': compressed textures (each requires its side-effect import)

When you need it

  • Signed CDN URLs: https://cdn.example.com/get?id=abc123 has no extension the loader can test against.
  • Blob or ObjectURL: URL.createObjectURL(blob) produces blob:... URLs with no extension.
  • Custom routing: /api/assets/hero-v2 where the server decides the content type.
  • Content-hashed paths without suffix: some build pipelines produce names like /static/abc123def instead of /static/abc123def.png.

If the URL *does* have an extension, you don't need parser; let auto-detection do its job. Only set parser when detection can't work.

loadParser is deprecated

The v7 loadParser field still works but emits a deprecation warning. Use parser for new code.

// Old (deprecated)
await Assets.load({ src: "...", loadParser: "loadTextures" });

// New
await Assets.load({ src: "...", parser: "texture" });

Topics

Every asset workflow is covered in a reference file. Pick the one that matches the question:

TopicReferenceWhen
Texture atlases and animationsreferences/spritesheet.mdLoading sprite sheets with AnimatedSprite
Video texturesreferences/video.md.mp4, .webm, autoplay, looping, mobile
Web and bitmap fontsreferences/fonts.md.woff2, .fnt, font families, SDF fonts
Animated GIFsreferences/gif.md.gif, GifSprite, playback control
Grouping assets by featurereferences/bundles.mdaddBundle, loadBundle, unloadBundle
Declaring everything upfrontreferences/manifests.mdAssets.init({manifest}) workflows
Cache lookups and cleanupreferences/caching.mdAssets.get, Assets.unload, Cache
Priming future assetsreferences/background.mdbackgroundLoad, backgroundLoadBundle
Loading screensreferences/progress.mdonProgress, LoadOptions progress
GPU-compressed formatsreferences/compressed-textures.md.ktx2, .basis, .dds, .ktx
Vector vs raster SVGreferences/svg.mdparseAsGraphicsContext, texture mode
Retina + format detectionreferences/resolution.md@{1,2}x, format preferences

Decision guide

  • Need to load a single image? Use Assets.load(url). No setup required.
  • Loading many assets grouped by level/scene? Use a bundle. See references/bundles.md.
  • Know all assets at build time? Use a manifest in Assets.init. See references/manifests.md.
  • Need a loading bar? Pass a progress callback to Assets.load. See references/progress.md.
  • Smooth transitions between levels? Background-load the next level. See references/background.md.
  • Memory budget matters? Use compressed textures and Assets.unload between screens. See references/compressed-textures.md and references/caching.md.
  • Need crisp SVG icons at any size? Load as Graphics, not texture. See references/svg.md.
  • Retina + WebP/AVIF? Configure texturePreference and use format patterns. See references/resolution.md.

Load options and error handling

There are two separate "options" concepts when loading assets:

  1. LoadOptions: the second argument to Assets.load/loadBundle. Controls error recovery, retries, progress, and completion callbacks across a whole load.
  2. data: a field on each asset descriptor. Forwards parser-specific options (scale mode, resolution, font family, autoplay flags, etc.) to the specific loader for that asset.

LoadOptions (per call)

await Assets.load(["hero.png", "enemy.png"], {
  onProgress: (p) => updateBar(p),
  onError: (err, url) => {
    const src = typeof url === "string" ? url : url.src;
    console.warn("failed:", src, err);
  },
  strategy: "retry",
  retryCount: 3,
  retryDelay: 250,
});
  • onProgress(progress): [0, 1] as assets in the call complete.
  • onError(error, url): url is string | ResolvedAsset. Guard before reading .src; when url is a string, .src is undefined.
  • strategy: 'throw' | 'skip' | 'retry' — default 'throw'. 'skip' resolves with any successful assets; 'retry' reattempts the failed ones.
  • retryCount — default 3, retries per asset when strategy is 'retry'.
  • retryDelay — default 250 ms between retries.

Global defaults live on Loader.defaultOptions, or pass loadOptions to Assets.init().

data options (per asset)

Each loader parser reads its own options from the data field on the asset descriptor. Use the table below to pick the right options for each asset type:

Asset typedata shapeKey optionsReference
Texture (image)TextureSourceOptionsresolution, scaleMode, alphaMode, autoGenerateMipmaps, antialias, addressModereferences/resolution.md
SVG{parseAsGraphicsContext?, resolution?}parseAsGraphicsContext for Graphics mode; resolution for sharper rasterreferences/svg.md
VideoVideoSourceOptionsautoPlay, loop, muted, playsinline, preload, updateFPS, crossorigin, mimereferences/video.md
Web fontLoadFontDatafamily, weights, style, display, unicodeRange, featureSettingsreferences/fonts.md
Bitmap font(none; auto-configured)Distance-field detection sets scale mode and mipmapsreferences/fonts.md
Spritesheet{texture?, imageFilename?, ignoreMultiPack?, textureOptions?, cachePrefix?}textureOptions forwards TextureSourceOptions (e.g. scaleMode) to the atlas image; texture to skip image load; imageFilename to override the referenced image; ignoreMultiPack to skip multi-pack follow-ups; cachePrefix to namespace framesreferences/spritesheet.md
GIFGifBufferOptionsfps, scaleMode, resolution, autoGenerateMipmapsreferences/gif.md
Compressed textureTextureSourceOptionsscaleMode, addressMode, autoGenerateMipmapsreferences/compressed-textures.md
JSON / Text(none)Returned as-is

Example combining LoadOptions and data:

await Assets.load(
  {
    alias: "hero",
    src: "hero.png",
    data: { scaleMode: "nearest", resolution: 2 },
  },
  { strategy: "retry", retryCount: 3 },
);

Inside a manifest or bundle, every entry can carry its own data:

await Assets.init({
  manifest: {
    bundles: [
      {
        name: "level1",
        assets: [
          { alias: "tiles", src: "tiles.png", data: { scaleMode: "nearest" } },
          { alias: "font", src: "hero.woff2", data: { family: "Hero" } },
          {
            alias: "clip",
            src: "intro.mp4",
            data: { autoPlay: false, muted: true },
          },
        ],
      },
    ],
  },
});

Runtime configuration

Assets.init(options) accepts, alongside basePath and manifest:

  • defaultSearchParams — string or Record<string, any> appended to every resolved URL. Useful for cache busting.
  • skipDetections: boolean — bypass browser format detection for faster init. Requires explicit texturePreference.format.
  • bundleIdentifier: BundleIdentifierOptions — customize how bundle keys resolve so the same alias can live in multiple bundles.
  • loadOptions: Partial<LoadOptions> — set the default strategy, retryCount, retryDelay, and callbacks for every subsequent Assets.load call.
  • preferences: Partial<AssetsPreferences>crossOrigin, preferWorkers, preferCreateImageBitmap, parseAsGraphicsContext.

After init, preferences can still be tuned:

Assets.setPreferences({
  crossOrigin: "anonymous",
  preferCreateImageBitmap: false,
});

for (const detection of Assets.detections) {
  console.log(detection.extension);
}

Assets.reset();
  • Assets.setPreferences(preferences) — push new preferences to every parser that supports them.
  • Assets.detections — getter exposing the registered FormatDetectionParser list; use when inspecting what formats the current environment advertises.
  • Assets.reset() — internal full reset (resolver + loader + cache). Intended for tests so a fresh Assets.init can run.

Common Mistakes

[CRITICAL] Using Texture.from(url) to load

Wrong:

const texture = Texture.from("https://example.com/image.png");

Correct:

const texture = await Assets.load("https://example.com/image.png");

In v8, Texture.from() only reads the cache. It does not fetch from a URL. Use Assets.load() first; the return value is the texture itself.

[HIGH] Using positional Assets.add signature

Wrong:

Assets.add("bunny", "bunny.png");

Correct:

Assets.add({ alias: "bunny", src: "bunny.png" });

The positional Assets.add(key, url) form was removed in v8. Use the options object with alias and src properties.

[HIGH] Not unloading textures between levels

Assets.load() caches textures indefinitely. For level-based games or screens with distinct asset sets, call Assets.unloadBundle() when transitioning to release GPU memory.

API Reference

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.76%
按下载量换算867

Claude

31.5%
按下载量换算809

Cursor

20.07%
按下载量换算515

Gemini CLI

9.25%
按下载量换算238

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills