Token导航 LogoToken导航TokenDH.com
效率敏感数据clawhub未标认证来源可访问clear审计通过

comic-drama-generate漫画生成

Agent Skill

comic-drama-generate 用于补充效率相关能力,适合在 OpenClaw 中需要让 Agent 承接效率相关任务时使用。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

10,160

周安装

411

GitHub Stars

公开资料未说明

下载量

3,189
OpenClaw

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

MIT-0

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:comic-drama-generate(漫画生成)
来源仓库:https://github.com/nanophotohq/comic-drama-generate
安装命令:
openclaw skills install comic-drama-generate
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install comic-drama-generate

简介

comic-drama-generate 用于生成具有很强角色一致性的连载漫画/漫剧剧集,提升创作效率。

  • 适用于需要多技能协作的漫画制作流程和系列内容创作场景。
  • 通过固定流程生成角色一致性强的漫画剧集内容。
  • 安装命令为 openclaw skills install comic-drama-generate,需注意多技能协同运行。
  • 建议核实具体触发条件和内容生成限制,避免过度消耗资源。

SKILL.md

name
comic-drama-generate
description
Generate serialized comic-drama / 漫剧 episodes through a fixed multi-skill production pipeline with strong character consistency. Use when the user wants to create AI 漫剧, 连载短剧, animated short drama, or episodic short-form storytelling; when they need a repeatable 三视图→关键帧→图生视频 workflow for consistent characters; when they want to turn an idea into scripts, storyboards, 15-second single-shot videos, and final edited episodes; or when they want to combine video-prompt-generator, nano-banana-pro, sora-2-generate, and local ffmpeg for storyboard-to-video production.

Comic Drama Generate

Use this skill to produce short comic-drama episodes through a deterministic, character-consistent pipeline.

Required prerequisite skills

Require these skills before using this skill:

  • video-prompt-generator
  • nano-banana-pro
  • sora-2-generate

If any are missing, install them first. Use this install pattern:

npx clawhub@latest install video-prompt-generator
npx clawhub@latest install nano-banana-pro
npx clawhub@latest install sora-2-generate

Before starting production, verify:

  • all three prerequisite skill folders exist
  • required API credentials are configured for the prerequisite skills
  • ffmpeg is available locally
  • there is enough credit/quota for script, image, and video generation

Credential note:

  • comic-drama-generate orchestrates other skills and does not introduce its own separate API key.
  • The prerequisite skills use their own environment-variable contract, typically NANOPHOTO_API_KEY.
  • Prefer normal sub-skill execution that relies on the prerequisite skill's configured env.
  • If direct script execution is necessary, first preserve the same env contract by ensuring NANOPHOTO_API_KEY is present in the shell.
  • Use explicit --api-key only as a fallback when the shell cannot inherit the expected env.
  • Do not treat direct reads from config files such as openclaw.json as the primary workflow; reserve them for debugging or recovery paths.

Mandatory production order

Always follow this exact order.

  1. Ask the user for style first

- Ask for visual style before generating anything. - Also confirm tone, audience, episode count, target platform, aspect ratio, and approximate episode duration. - Lock the series bible first: premise, character roster, style keywords, and production constraints.

  1. Use video-prompt-generator to generate the script foundation

- Generate the story premise, dramatic arc, episode beats, and scene-level writing foundation. - Make the episode structure suitable for short-form serialized漫剧. - Keep the pacing compatible with downstream 15-second shot generation.

  1. Use nano-banana-pro to generate character turnarounds

- Generate character three-view sheets for all core recurring characters. - Treat the three-view sheets as the canonical identity source for face, hair, silhouette, wardrobe, and palette. - Do not skip this step when character consistency matters.

  1. Use nano-banana-pro to generate keyframes

- Generate keyframes for each planned shot using the turnaround image URLs as reference inputs. - For multi-character shots, include all relevant turnaround URLs. - Use keyframes to lock composition, costume continuity, environment cues, and emotional staging.

  1. Use video-prompt-generator to write the storyboard / shot script

- Convert each episode into shot-level prompts and a shot list. - Write for single-shot generation with clean handoff into video prompts. - Leave a hook at the end of every episode.

  1. Use sora-2-generate to create 15-second single-shot videos

- Default to 15-second shots unless the user explicitly wants a different supported duration. - Prefer imageToVideo for character shots. - Pass keyframe URLs as image inputs. - Use textToVideo only for shots where character consistency is unimportant, such as atmosphere inserts or environment transitions.

  1. Use local ffmpeg to trim, arrange, and merge clips

- Download and save each finished shot locally. - Trim or normalize clips locally if needed. - Merge shots in deterministic order with ffmpeg. - Export final episode cuts under a dedicated project directory.

Non-negotiable working rules

  • Prefer this consistency chain at all times:

- character turnarounds → keyframes → image-to-video

  • Use public URLs, not local file paths, when passing assets between:

- character turnarounds → keyframes - keyframes → video generation

  • Do not pass local files into nano-banana-pro or sora-2-generate APIs for these handoff steps.
  • Do not skip directly from text prompt to final video for main character shots unless the user explicitly accepts weaker consistency.
  • Keep exact turnaround and keyframe URLs recorded in project docs because downstream generation depends on them.
  • Rewrite storyboard timing before generation if model limits, quota, or rate limits make the original plan unrealistic.

Story and pacing guidance

  • Design shots around the active video model's supported duration.
  • For this workflow, prefer a shot grammar that fits 15-second single-shot clips.
  • If an episode needs more time, split it into more shots rather than overloading one prompt.
  • Ensure each episode ends with a hook, reveal, reversal, question, or emotional cliffhanger.

Suggested project structure

project/
  docs/
    series-bible.md
    episodes.md
    shot-list.md
    asset-urls.md
  assets/
    characters/
    keyframes/
  shots/
    ep01/
    ep02/
  edits/
  final/

Execution notes

  • Use video-prompt-generator twice when needed:

- once for premise / episode writing - once for shot-level prompt refinement

  • Use nano-banana-pro for both:

- character turnaround generation - keyframe generation

  • Use sora-2-generate mainly in imageToVideo mode for character-led shots.
  • Use local ffmpeg for final editorial assembly, not remote editing tools.

Expected deliverables

Produce organized outputs that make the project easy to resume:

  • docs/series-bible.md — premise, style, characters, world rules
  • docs/episodes.md — episode summaries and hook endings
  • docs/shot-list.md — shot-by-shot structure for each episode
  • docs/asset-urls.md — turnaround URLs, keyframe URLs, and video URLs
  • local shot files under shots/
  • local edit outputs under edits/
  • final episode exports under final/

Read these references when needed

  • references/workflow.md — end-to-end production workflow
  • references/install-checklist.md — prerequisite skill verification and install steps
  • references/asset-rules.md — character turnaround, keyframe, and image-to-video consistency rules

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

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

补充不同宿主或平台的使用分布数据

能力 5

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

OpenClaw

98.75%
按下载量换算3,149

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。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