Communitygithub.com

Pluviobyte/rnskill

Creates sequential keyframes for visual references of power manifestations, avatar projections, or character force displays, then uses user-selected or available image-to-video tools to generate, inspect, and composite action videos. Does not bind to any specific video platform.

What is rnskill?

This skill focuses on breaking down a single readable power display (like a colossal avatar or domain expansion from Chinese web novels) into sequential states, and then assigning the motion between adjacent states to video generation tools for segmented creation, inspection, and final compositing. It is platform-independent, meaning it doesn't default to any particular image or video generation service; instead, it adapts to the user's preferred tools or the most suitable available option based on constraints like capability, budget, and output requirements. The skill meticulously manages character identity, spatial relationships (e.g., the user's body in the foreground with the giant avatar behind), and action continuity using a structured continuity table, preventing common errors like mirrored poses or conflicting action phases. It can receive event cards from 'novel-distill' or work directly from user ideas, and it clearly distinguishes original text facts from visual interpretations and adaptations, ensuring that visual references do not pollute the canonical story facts. The final deliverable is a set of independent keyframes and/or a downloadable composite video, along with a detailed record of prompts, tools used, and any residual deviations.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/Pluviobyte/rnskill/tree/HEAD/skills/avatar-keyframe-video

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Giant Avatar Consecutive Frame to Video

Break a readable power display into sequential states, then hand off the motion between adjacent states to video generation. Deliver a plan, keyframes, or the final video as per the user's current request; do not automatically expand the scope.

Picking up from Current Progress

  • Approved style reference: Use directly as the master template; do not re-select styles.
  • Approved consecutive frames: Check for contradictions between frames before generating video; do not re-generate images.
  • Only plan, skill, prompts, or keyframes needed: Complete the corresponding part; do not automatically submit a paid video task.
  • When the user says "continue after selection," stop at the selection point; if already selected or authorized, do not reconfirm.
  • If a non-critical preference is missing and autonomous decision is allowed, adopt a reasonable default and proceed. Rainy night, eight frames, 25 seconds, a sword with music are not fixed requirements.

Can accept event cards from novel-distill, or directly use the user's ideas; no additional skill installation required. Distinguish between original text facts (F), audiovisual processing (P), and new adaptations (A). Visual references like domains, avatars, and weapons from reference videos must not be backfilled as novel facts. When no original text exists, refer to it as a visual concept interpretation.

Tool Selection: Platform Agnostic

This skill has a fixed production method but no default final platform. Image generation, image-to-video, and editing/mixing can be done with different tools or within the same platform.

  • If the user specifies a tool or has an existing project, prioritize using it, verifying its actual capabilities. Do not automatically choose Pexo just because it was used in a historical case.
  • If unspecified, select based on currently available tools, reference image constraint capabilities, duration, downloadable output, and budget, providing a brief rationale. If a viable plan already exists and the user permits autonomous decisions, no additional confirmation step is needed for platform selection.
  • If no usable generative tools are available, deliver independent keyframes, a general segment brief, and prompts, explaining the missing capabilities. Do not describe prepared materials as already generated video.
  • First verify the current API or interface, then map the general brief into specific parameters. Platform names do not equal model capabilities; do not assume every platform supports multi-image reference, start/end frames, or native sound effects.

1. Locking the Co-Frame Relationship and Continuity

Use a brief continuity table to record existing information:

ItemContent to Lock
CharactersFace, age, hairstyle, robe silhouette, material, identifying features
AvatarDefault is the user's magnified projection, same face and attire; if the user specifies a deity/alien form, record the correspondence
Co-frame spaceUser's body in the foreground, avatar behind, taller than buildings; both the small figure's silhouette and the giant's face readable simultaneously
Body and frame coordinatesWhich anatomical hand holds what, where the start and end points are in the frame
SceneArchitectural landmarks, light sources, weather, camera position and action axis
Material evolutionE.g., silvery phantom → skin tone cloth → edge dissipation; do not impose if the user does not need it
Attack executionStart → charge → release → recovery, including number of times, direction, affected objects, and environmental reactions

Map the same action, not a horizontal mirror. The user's anatomical right arm corresponds to the avatar's right arm; do not write ambiguous "mirror pose" causing left-right reversal. The avatar must not become a second opponent in the distance; a close-up of the face alone cannot replace the required co-frame key image.

Default to a single camera position or slight push-in based on the master template to reduce continuity issues; when the user wants multiple camera positions, clearly define the cuts and action axis, do not turn a single position into a style prohibition.

2. Design State Changes First, Then Generate Independent Frames

Read references/keyframes.md. For each frame, record the user's pose, avatar pose/transparency, weapon state, environmental reaction, differences from the previous frame, and actions not yet taken. Differences should be visually present in the image.

A common structure is "charge → domain first opens → first appearance → expansion → consolidation → gather weapon and raise hand → simultaneous release → aftereffects," which can be merged or split. When the gap between "gather weapon" and "raise hand" is too large, add transition states; do not cram more requirements into the same prompt.

Use currently available image generation/editing tools; when the user specifies a platform or model, confirm the actual callable capabilities. If the API does not return a version, record the channel and "version undisclosed"; do not claim specific model names like GPT-Image 2.

Generate one independent frame at a time, maintaining the aspect ratio. Usually input "approved master + last approved frame": the master maintains identity, the previous frame maintains state. Correct errors in the previous frame first; do not propagate errors, and do not input all images at once. Early frames must suppress climax elements from the master: no giant face if not yet manifested, no complete sword light if not yet released.

Inspect each frame after generation before continuing. If only the avatar's arm is wrong, make a targeted correction, preserving the correct user body, face, clothes, and scene. Reasonable differences in local textures are allowed; major contradictions in identity, hand side, and action phase must be resolved first.

3. Assigning Inter-Frame Motion to Video Segments

Read references/video-and-qa.md. Segment based on duration capability and action complexity, with adjacent segments sharing a passed boundary state. Shared images do not guarantee natural continuity after generation.

For each segment, write the initial state, final state, process reference, main action, camera movement, correspondence between parties, environmental forces, sound beats, and ending pose. Images fix appearance and state; video prompts describe the motion between states; do not create a static image slideshow.

If there are dedicated start/end frame fields, use them according to the API; a plain image_list can only be called an ordered reference. When there is no end frame constraint, check the actual segment end; if necessary, use the passed segment end as the next segment's start frame. Only submit independent segments simultaneously when boundaries are clear and the tool supports the corresponding constraints.

Only read the optional adaptation guide references/pexo.md when Pexo is actually used this time. It is not the default entry or required dependency; for other tools, follow the general brief and current capabilities.

4. Inspection, Correction, and Composition

First, look at the actual generated material, then decide on regeneration. Prioritize fixing co-source relationships, hand side and direction, repeated actions, action regression, and major identity drift. Judge material deviations against the original target; a platform's auto-check text alone cannot trigger a paid retry.

Co-frame does not mean synchronized. If the user has sheathed their weapon while the avatar still holds it up, that is a critical error. Check the period before and after each release and every junction point; use reasonable cropping, matching cut points, and sound effect bridges to handle minor differences; do not use reverse playback or long dissolves to mask erroneous poses.

Record original and corrected versions; do not overwrite the only successful piece of material. Clearly define defects in each round; reserve retry and post-processing budget. If the user sets an upper limit, do not exceed it without permission; also report deviations from the assistant's own cost targets. Multiple corrections should be itemized, not lumped into a single "round" that incorrectly implies free retries.

Arrange sound according to charge, release, and aftereffects; preserve effective ambient and action sound effects. If the soundtrack is too long, cut and fade it out first; only consider regeneration if the content is unsuitable. "Replacing the audio track" cannot be claimed as "overlaying and preserving." If the platform cannot mix audio, external post-processing can be done under existing authorization, but clearly explain the division of labor.

Local technical inspection example:

python3 scripts/audit_video.py /absolute/path/final.mp4 --cuts 6.04,12.08,17.12 --out /absolute/path/qa

Cut points must be actual timestamps. Requires Python 3, ffmpeg, ffprobe. The script outputs media parameters, full timeline overview, extracted frames at cut points, black frame detection, and volume checks; it does not judge character synchronization, subjective audio quality, or presence of dialogue; those must be checked manually and via appropriate audio inspection.

Deliverable

Deliver independent keyframes and/or a downloadable final video according to the agreed scope, preserving the state table, prompts, and production records. Record the actual tools/confirmed model, timeline, retries, costs, and residual deviations. For the final video, also record source material and output specifications; do not claim upscaling as native high resolution, and do not mistake platform "completion" for acceptance.

Ultimately, directly present the work and any significant limitations. When real examples are needed, read references/worked-example.md; historical prices and results do not constitute a promise of success rate, budget, or follower growth.

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

Pluviobyte/rnskill

口播基础素材包生成。转录口播、识别口误、生成审核页;用户确认后剪出新视频,Agent 再基于剪后视频重新转写、AI 校对字幕,输出后续口播成片可用的 source_cut.mp4 和 subtitles.srt。触发词:剪口播、处理口播素材、准备口播素材、识别口误、基础素材包

Pluviobyte/rnskill

Create Ian Xiaohei-style HTML/SVG motion illustrations for Chinese articles, scripts, storyboard scenes, workflow explanations, concept metaphors, and short video visual aids. Use when the user asks for 小黑 SVG、漫画感 HTML、分层 SVG 动效、把小黑生图做成 HTML/SVG、可编辑矢量动效、口播流程动画、手绘漫画动效、正文配图动画, or wants a Xiaohei illustration style adapted into controllable HTML/SVG/GSAP output instead of raster image generation.

Pluviobyte/rnskill

口播视频成片 Skill。把文章/口播稿/SRT、剪后视频和 HTML/图片素材串成分镜稿、时间线预览和最终 MP4;成片比例和动画风格从用户配置读取,动画默认使用小黑风格。触发词:口播成片、做分镜稿、时间线预览、合成口播视频、导出竖屏MP4

Pluviobyte/rnskill

自进化 skills。记录用户反馈,更新方法论和规则。触发词:更新规则、记录反馈、改进skill

Pluviobyte/rnskill

dontbesilent 目标清晰化。用维特根斯坦的语言哲学把模糊的目标审计成可检查的交付物。 触发方式:/dbs-goal、/目标、「帮我搞清楚目标」「我想做个人 IP」「我的目标是成为...」「我想变得更...」 Goal clarification using Wittgenstein's philosophy of language. Audits fuzzy goals into checkable deliverables. Trigger: /dbs-goal, "help me clarify my goal", "I want to become...", "my goal is...

Pluviobyte/rnskill

dontbesilent 短视频开头优化。诊断开头问题 + 生成优化方案。 触发方式:/dbs-hook、/hook、「帮我优化开头」「开头怎么写」 Short video opening optimization with diagnosis and solutions. Trigger: /dbs-hook, "optimize my opening", "how to write opening

Pluviobyte/rnskill

把当前诊断的关键状态存到本地,下次回来可以接着用。 触发方式:/dbs-save、/存档、「保存这次诊断」「记下来」「这个结论留着」 Save the current diagnosis state to disk for cross-session recall. Trigger: /dbs-save, "save this diagnosis", "remember this

Pluviobyte/rnskill

dontbesilent 商业工具箱主入口。双模式:任务前路由(你的问题该用哪个 skill)+ 任务后导航(刚做完诊断,下一步该干什么)。 触发方式:/dbs、/商业、「帮我看看」、「下一步怎么走」 Main entry point for dontbesilent business toolkit. Dual mode: pre-task routing + post-task navigation. Trigger: /dbs, "help me with my business", "what's next

Pluviobyte/rnskill

孙宇晨.skill 的别名。触发后立刻读取并执行 `../孙宇晨/SKILL.md`,不单独维护内容。 触发:孙割.skill、/孙割、"用孙割视角看"、"孙割会怎么处理"、"切换到孙割"。

Pluviobyte/rnskill

蒸馏孙宇晨(孙割)公开可查的决策模式,做成一套非扮演的决策分析工具:拆解他如何把注意力、身份、叙事当成可购买、可叠加、可兑换的资产来经营,用来给你自己的选题/发布/公关/危机决策做沙盘推演,不用于替他或任何真实人物发声、编造言论。 触发:孙宇晨.skill、孙割.skill、/孙宇晨、/孙割、"用孙割视角看"、"孙宇晨会怎么做"、"孙割会怎么处理"、"注意力收割模型"、"蒸馏孙宇晨"。

Pluviobyte/rnskill

将参考图或简要描述解码为可编辑的半色调纸张拼贴规范,默认使用 Codex 内置 image_gen 生成品牌一致的静帧和透明图层,再用本地 FFmpeg 或 HyperFrames 制作“从空背景逐件组装”的确定性动态视频。用户提到半色调拼贴、纸张拼贴、剪纸拼贴、编辑式拼贴、杂志拼贴、复古印刷拼贴、拼贴动效、拼贴动画、剪纸动画、纸片组装、逐层组装、从无到有、元素飞入拼装、定格拼贴、定格动画、Arcads 风格、参考图拆解、风格拆解、反向拆解视觉、照这个风格做、做类似的、生成拼贴静帧、让这张图动起来、让这些元素组装起来、制作拼贴广告或拼贴解释视频时使用。英文触发语义包括 editorial collage、halftone collage、paper collage、paper-cut motion、stop-motion graphic、assemble from empty、build-on motion、analyze this、break this down、reverse-engineer this look、generate the stills、animate this、make these move。默认不调用 Arcads、Nano Banana、Seedance 或 Kling;只有用户明确要求且对应连接器可用时才进入外部生成路径。

Pluviobyte/rnskill

Create deterministic editorial-style cover images with a warm paper-gray background, oversized black slab-serif/Chinese display typography, generous whitespace, and a dotted vector analysis icon. Use when the user asks for a “点阵编辑风封面”, “暖灰杂志风封面”, “Codex + Hyperframes 同款封面”, or wants to reproduce/adapt the specific minimal left-text/right-icon banner style. Produce editable SVG and a rendered PNG; do not use this skill for generic covers without this visual direction.

Pluviobyte/rnskill

Chinese AI/tech writing de-AI editor for posts, X/Twitter threads, technical essays, product notes, model reviews, and public-writing drafts. Use when the user asks to 去AI味, 改得像本人, 写推特post, 精修中文AI技术文章, or complains about AI-flavored shells such as 不是A而是B, 真正/其实/本质上, 更重要的是, 冒号讲义腔, 空泛总结, 顺滑但没作者判断的稿子.

Pluviobyte/rnskill

内容复盘工作流:取数 → 爆款分级(R/M/Tier)→ 按需拉取内容本体 → 归因 → 沉淀(方法论/金句库/案例库/概念库/课程灵感)+ 选题卡回写。 Use when the user says 复盘xx, 内容复盘, xx数据怎么样, xx表现如何, 复盘一下这条, 为什么这条爆了, 为什么这条没爆. 沉淀铁律:复盘不沉淀,等于白复盘——每次至少一条资产入库,或写明为什么没有。

Pluviobyte/rnskill

实操长片策划工作流(B站/YouTube 新赛道,实操+真人出镜)。产出可直接照着录的策划稿:测试题组(含可粘贴 prompt)、 结构时间轴(全身出镜段 vs PiP 段)、出镜口播稿、录屏操作清单、悬念与彩蛋设计。 Use when the user says 做实操长片, 实操视频, 实测视频, 对比实测, xx做成实操, 出个实测策划, 这个选题拍实操, or ra-选题 routes a 实操长片 立项 here. 策划稿落 `01-内容生产/视频工作台/制作中/<日期-主题>/策划稿.md`,frontmatter status 是与制作端交接的唯一状态机。

Pluviobyte/rnskill

视频脚本洗稿和二创脚本工作流。Use when the user provides an existing transcript, oral draft, viral script, competitor video text, rough topic notes, or wants 洗稿, 二创改写, 爆款视频脚本重写, 逐字稿改成短视频脚本, rewrite with ra-人话, and queue the approved script into 01-内容生产/视频工作台/待制作/ as the handoff contract for video production. If the user provides a video URL or local video file and asks 视频洗稿、视频二创、洗稿并制作视频, use ra-video-wash-pipeline first.

Pluviobyte/rnskill

选题全生命周期调度:记录选题、深化选题、选题推荐、立项分发(视频/图文/文章/实操长片/多形态)、发布归档。 Use when the user says 灵感xx, 想法xx, 记录选题, 这个可以做个选题, 先存着这个链接, 深化选题/深化一下xx, 给我N个选题, 帮我出选题, 选题推荐, 立项, 做成视频, 制作视频, 做成文章, 制作文章, 做成图文, 制作图文, 做成实操长片, 图文视频文章都做/都分发, 我要自己出镜讲, xx发布了, xx要改; or when the user gives a link plus a target form (链接+制作视频/文章/图文). Owns the 00-选题池 card system: one topic = one card; user never moves files — this skill does.

Pluviobyte/rnskill

短视频钩子(开头)类型选择与写法,中文名「钩子」。Use when 起钩子、写钩子、钩子怎么写、钩子没思路、设计前5秒钩子、写视频开头/口播稿开头, or when the user asks 视频开头怎么做、开头模板、hook 类型、帮我起个钩子, including during ra-洗稿 script rewriting and 选题深化. Provides a scenario-based taxonomy of 7 hook types with decision flow, prerequisites, sentence templates, and common mistakes. For diagnosing and optimizing an existing hook draft, use dbs-hook instead; this skill chooses the type, dbs-hook polishes the execution.

Pluviobyte/rnskill

End-to-end video production orchestration for the content-creation workspace. Use when the user asks to make, recreate, package, render, QC, or archive a video; when the user says 制作待制作队列, 按交接稿制作, 制作队列视频, or points at a handoff project folder under 01-内容生产/视频工作台/待制作/ (contract file 交接稿.md inside) or an image-card handoff under 01-内容生产/图文工作台/待制作/ (its frontmatter is the binding production contract); when a request could route to HyperFrames, Remotion, captions, TTS, download/transcription, website/product videos, motion graphics, or reference replication; or when the user needs production planning before editing.

Pluviobyte/rnskill

Create dark cinematic SaaS/product motion videos in the Presenton-like "magic UI" style: black starfield stage, bottom purple glow, large kinetic type, cyan-to-magenta CTA buttons, prompt cards, floating UI templates, integration badges, model rings, export objects, speed-blur transitions, and short product capability arcs. Use when the user asks for "这种风格", "同款风格", dark SaaS promo, AI product demo, magic UI product video, or a style-level creation inspired by a reference video. Do not use for pixel-level or frame-accurate replication; use rn-replica-qc first for that.

Related Skills