Token导航 LogoToken导航

Agent Skill · 图像处理

pixverse-ai-image-and-video-generator(pixverse ai 图像和视频生成器)

用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。它适合让 Agent 根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。使用时需要确认输入图片、版权来源、输出格式和模型限制;涉及人物、品牌、商品或公开展示素材时,应额外核对授权、真实性和内容合规边界。

pixverse-ai-image-and-video-generator 是图像处理类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证信息还不明确,第三方扫描有提醒项,按安装量在图像处理类的 1,562 个 Skill 里排第 477。

总安装 6,022周安装 246GitHub Stars 45下载 1,929核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:pixverse-ai-image-and-video-generator(pixverse ai 图像和视频生成器)
来源仓库:https://github.com/pixverseai/skills
仓库路径:skills/pixverse-ai-image-and-video-generator
安装命令:
npx skills add https://github.com/pixverseai/skills --skill pixverse-ai-image-and-video-generator
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/pixverseai/skills --skill pixverse-ai-image-and-video-generator

这个 Skill 是做什么的

图像处理类 Skill 处理生成、裁剪、压缩和批量转换图片的任务。

用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。适合根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。通过 npx skills add 命令安装,需结合原始 README 核验具体用法和功能细节。使用时需确认输入图片、版权来源、输出格式和模型限制,涉及人物或品牌时需核对授权与合规边界。

pixverse-ai-image-and-video-generator 属于图像处理类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/pixverse-ai-image-and-video-generator

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/pixverseai/skills --skill pixverse-ai-image-an…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

pixverse-ai-image-and-video-generator 怎么安装?

在终端执行 npx skills add https://github.com/pixverseai/skills --skill pixverse-ai-image-and-video-ge…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

pixverse-ai-image-and-video-generator 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

pixverse-ai-image-and-video-generator 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

pixverse-ai-image-and-video-generator 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

pixverse-ai-image-and-video-generator 用的人多吗?

公开数据是累计安装 6,022 次,近一周安装 246 次,下载 1,929 次,GitHub 45 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

pixverse-ai-image-and-video-generator 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理图像处理类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

图像处理类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
ai-image-generation

AI 图像生成 · Codex、Claude

4,365,552382调用外部服务未明确有提醒
openai-image-gen

OpenAI 图像生成 · OpenClaw

1,906,88432需要联网MIT-0有提醒
image

图像 · OpenClaw

525,99926需要联网MIT-0通过
baoyu-cover-image

宝玉封面图 · Codex、Claude

474,24016,779只读MIT有提醒
baoyu-image-gen

宝玉图像生成器 · Codex、Claude

468,00716,778接触敏感数据MIT通过
pixverse-ai-image-and-video-generator

pixverse ai 图像和视频生成器 · 本页

6,02245操作浏览器未明确有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

PixVerse CLI — Master Skill

What is PixVerse CLI

PixVerse CLI is the official command-line interface for PixVerse — an AI-powered creative platform for generating videos and images. It is essentially a UI-free version of the PixVerse website: all features, models, and parameters are aligned with the web experience at app.pixverse.ai.

It is designed for:

  • AI agents (primary) — structured JSON output, deterministic exit codes, and pipeable commands for autonomous workflows (Claude Code, Cursor, Codex, custom agents)
  • Developers & power users — scriptable video/image generation without leaving the terminal
  • Automation — batch processing, CI/CD pipelines, content production workflows

Key facts:

  • Generating content consumes credits from the user's PixVerse account (same pricing as the website)
  • Only subscribed users can use the CLI — see subscription plans
  • All output can be returned as structured JSON via --json flag
  • English only

Installation

npm install -g pixverse

Or run without installing:

npx pixverse

Verify:

pixverse --version

Requires Node.js >= 20.


Quick Start

# 1. Install
npm install -g pixverse

# 2. Authenticate (OAuth device flow — opens browser)
pixverse auth login --json

# 3. Create a video (waits for completion by default)
RESULT=$(pixverse create video --prompt "A cat astronaut floating in space" --json)
VIDEO_ID=$(echo "$RESULT" | jq -r '.video_id')

# 4. Download the result
pixverse asset download $VIDEO_ID --json

To skip waiting and poll later:

RESULT=$(pixverse create video --prompt "A cat astronaut floating in space" --no-wait --json)
VIDEO_ID=$(echo "$RESULT" | jq -r '.video_id')
pixverse task wait $VIDEO_ID --json
pixverse asset download $VIDEO_ID --json
Windows users: For a full PowerShell pipeline example (T2I → I2V → sound → upscale → download), see skills/examples/windows/powershell-text-to-video.ps1.

Authentication

PixVerse CLI uses OAuth device flow — no need to manually copy tokens:

  1. Run pixverse auth login --json
  2. The CLI prints an authorization URL
  3. Open the URL in your browser and authorize
  4. The token is stored automatically in ~/.pixverse/

Details:

  • Token is valid for 30 days
  • CLI sessions are independent from your web/app sessions
  • If token expires (exit code 3), re-run pixverse auth login --json
  • Run pixverse auth status --json to check login state and credits

Capabilities Overview

I want to...Use skill
Create a video from text or imagepixverse:create-video
Enhance a video prompt for better resultspixverse:prompt-enhance
Edit video content with AI (replace subjects, swap outfits, change backgrounds)pixverse:modify-video
Animate a character with motion from a reference videopixverse:motion-control
Create or edit an imagepixverse:create-and-edit-image
Extend, upscale, or add audio to a videopixverse:post-process-video
Create transition animation between framespixverse:transition
Check generation progresspixverse:task-management
Browse, download, upload, or delete assetspixverse:asset-management
Organize assets into named folderspixverse:saved-folders
Set up auth or check accountpixverse:auth-and-account
Browse and create from effect templatespixverse:template
Manage workspaces (list, switch, status)pixverse:workspace
Generate Mondo-style posters and coverspixverse:mondo-poster-design
Looking up models or parameters? Don't wait until you're generating — read the relevant capabilities file directly: - Video models & constraints → skills/capabilities/create-video.md (Model Reference section) - Image models & constraints → skills/capabilities/create-and-edit-image.md (Model Reference section)

Model Quick Reference

Use this to pick a model before diving into a sub-skill.

Video Models (pixverse create video --model <value>)

Model--model valueMax QualityDuration
PixVerse V6 *(default)*v61080p1–15s
PixVerse C1pixverse-c11080p1–15s
PixVerse v5.6v5.61080p1–10s
Sora 2sora-2720p4 8 12s
Sora 2 Prosora-2-pro1080p4 8 12s
Veo 3.1 Standardveo-3.1-standard1080p4 6 8s
Veo 3.1 Fastveo-3.1-fast1080p4 6 8s
Veo 3.1 Liteveo-3.1-lite1080p4–6s
Grok Imaginegrok-imagine720p1–15s
Seedance 2.0 Standardseedance-2.0-standard1080p4–15s
Seedance 2.0 Fastseedance-2.0-fast720p4–15s
Kling O3 Prokling-o3-pro720p3–15s
Kling O3 Standardkling-o3-standard720p3–15s
Kling 3.0 Prokling-3.0-pro720p3–15s
Kling 3.0 Standardkling-3.0-standard720p3–15s

Image Models (pixverse create image --model <value>)

Model--model valueMax Quality
Qwen Image *(default)*qwen-image1080p
GPT Image 2gpt-image-2.02160p
Seedream 5.0 Liteseedream-5.0-lite1800p
Seedream 4.5seedream-4.52160p
Seedream 4.0seedream-4.02160p
Gemini 2.5 Flash (Nanobanana)gemini-2.5-flash1080p
Gemini 3.0 (Nano Banana Pro)gemini-3.02160p
Gemini 3.1 Flash (Nano Banana 2)gemini-3.1-flash2160p
Kling Image O3kling-image-o32160p
Kling Image V3kling-image-v31440p

For full parameter constraints (aspect ratios, quality per model, mode support), read the capabilities files listed above.


Workflow Skills

I want to...Use skill
Generate video from text end-to-endpixverse:text-to-video-pipeline
Animate an image into videopixverse:image-to-video-pipeline
Generate image then animate itpixverse:text-to-image-to-video
Iteratively edit an imagepixverse:image-editing-pipeline
Modify a video and enhance itpixverse:modify-video-pipeline
Full video production (create + extend + audio + upscale)pixverse:video-production
Animate a character with a motion referencepixverse:motion-control-pipeline
Create multiple items in parallelpixverse:batch-creation
Generate a Mondo-style poster end-to-endpixverse:mondo-poster-pipeline
Generate poster then animate into videopixverse:mondo-poster-to-video-pipeline
Storyboard → 4-shot video from a single promptpixverse:storyboard-to-video

Reference Materials

Located in skills/references/. These are read-only knowledge bases that capabilities and workflows draw from — no CLI commands, just curated design knowledge.

ReferencePathContent
Mondo Artist Stylesreferences/mondo-poster/artist-styles.md37 artist styles with prompt keywords across 7 categories
Mondo Composition Patternsreferences/mondo-poster/composition-patterns.md8 composition techniques (negative space, silhouette, geometric framing, etc.)
Mondo Genre Templatesreferences/mondo-poster/genre-templates.mdGenre-specific prompt templates for film, book covers, and album covers

All Commands

CommandDescription
auth loginLogin via browser (OAuth device flow)
auth statusCheck authentication status
auth logoutRemove stored token
create videoText-to-video or image-to-video
create imageText-to-image or image-to-image
create transitionCreate transitions between keyframes
create speechAdd lip-sync speech to video
create soundAdd AI sound effects to video
create modifyModify video content with a prompt at a keyframe
create extendExtend video duration
create upscaleUpscale video resolution
create referenceGenerate video with character references
create motion-controlGenerate video with character image + motion reference video
create templateCreate video or image from an effect template
template categoriesList template categories
template listBrowse templates (with optional category filter)
template searchSearch templates by keyword
template infoGet template details
task statusCheck task status
task waitWait for task completion
asset listList generated assets (with --source and --off-peak filters)
asset infoGet asset details
asset downloadDownload a generated asset
asset uploadUpload a local file or HTTPS URL to asset library
asset deleteDelete an asset
saved listList saved folders
saved itemsList items in a saved folder
saved newCreate a new saved folder
saved renameRename a saved folder
saved addAdd assets to a saved folder
saved removeRemove assets from a saved folder
saved deleteDelete a saved folder
account infoView account info and credits
account usageView credit usage records
workspace listList all workspaces
workspace statusShow currently active workspace
workspace switchSwitch to a different workspace
workspace manageOpen workspace management in browser
subscribeOpen subscription page in browser
config listList all config values
config getGet a config value
config setSet a config value
config resetReset config to defaults
config pathShow config file path
config defaultsManage per-mode creation defaults

Global Flags

FlagDescription
--json or -pPure JSON output to stdout (required for agent use)
--workspace-id <id>Per-command workspace override (0 = personal). Not persisted — only affects the single invocation.
-V, --versionShow CLI version
-h, --helpShow help for any command

Every command supports --json. All examples in skills use --json for machine-readable output.

Interactive mode: Run any creation command without arguments (and without --json) to enter the interactive wizard.


Output Contract

JSON mode (--json)

  • stdout: Pure JSON only. No spinners, no progress text, no decorative output.
  • stderr: All errors, warnings, and diagnostic messages — including error payloads in --json mode (as of CLI v1.1.4: pixverse task … and pixverse template … errors are also routed to stderr, preserving the stdout-is-success contract).
  • Parse stdout with jq or any JSON parser.

Universal JSON fields

All --json object payloads (both success on stdout and errors on stderr) automatically include:

FieldWhen presentMeaning
trace_idWhenever the command made an HTTP request and the API returned an Ai-Trace-Id headerUpstream request id — include this when reporting bugs. Array / primitive payloads are not augmented.
codeError payloads from API failuresBackend error code (from ApiError). Pair with trace_id for support.
errorError payloadsHuman-readable error message.
cost_creditscreate … success payloads, only when backend returns a positive integerCredits charged for this creation request. Absent when the API returns 0, null, or omits the field.

In interactive (non-JSON) text mode, cost_credits surfaces as Cost: N credits after Submitted!.

Exit Codes

CodeNameMeaningRecovery
0SUCCESSCompleted—
1GENERAL_ERRORUnexpected errorCheck stderr for details
2TIMEOUTPolling timed outIncrease --timeout or use --no-wait then pixverse task wait
3AUTH_EXPIREDToken invalid/expiredRe-run pixverse auth login --json
4CREDIT_INSUFFICIENTNot enough creditsCheck pixverse account info --json, wait for daily reset or upgrade
5GENERATION_FAILEDGeneration failed/rejectedCheck prompt, try different parameters
6VALIDATION_ERRORInvalid parametersCheck flag values against enums in each skill

Workspace error auto-recovery

When a request fails because the active workspace is no longer accessible (e.g. user was removed from a team), the CLI automatically resets to personal workspace (ID=0) and asks you to retry. This does not trigger when --workspace-id override is active or the failing request is a workspace management command.

Error handling pattern

RESULT=$(pixverse create video --prompt "A sunset over mountains" --json 2>/tmp/pv_err)
EXIT_CODE=$?

if [ $EXIT_CODE -eq 0 ]; then
  VIDEO_ID=$(echo "$RESULT" | jq -r '.video_id')
  echo "Success: $VIDEO_ID"
  pixverse asset download $VIDEO_ID --json
elif [ $EXIT_CODE -eq 3 ]; then
  echo "Token expired, re-authenticating..."
  pixverse auth login --json
elif [ $EXIT_CODE -eq 4 ]; then
  echo "Not enough credits"
  pixverse account info --json | jq '.credits'
elif [ $EXIT_CODE -eq 5 ]; then
  echo "Generation failed — check prompt or parameters"
  cat /tmp/pv_err
else
  echo "Error (code $EXIT_CODE)"
  cat /tmp/pv_err
fi

适合场景

01

文本生成图片

02

图片风格化

03

产品图和创意图

04

需要 FLUX 模型时

能力概览

能力 1

调用 FLUX 图像模型

能力 2

支持文本生图和图像改写

能力 3

覆盖 LoRA 或风格适配

能力 4

适合创意视觉生成

平台分布

Codex

34.97%
按下载量换算675

Claude

32.84%
按下载量换算633

Cursor

20.77%
按下载量换算401

Gemini CLI

8.91%
按下载量换算172

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills