Cat Workday Vlog · Open Skill 公開スキル
Cat Workday Vlog
**Cat Workday Vlog · One Cat, One Job, One Cute Cut · ref2v Native · Foley Kept** *The agent skill for creators who want a finished **workplace-comedy short starring their cat**—not a human in a cat filter, a still portrait, or a first-frame freeze of a lookbook plate* **Cat Workday Vlog** is a production-native Agent Skill for Digen. It orchestrates a complete **reference-to-video (ref2v)** pipeline for one cat workday episode: **Idea (or your cat photo) → job lock → identity plate + empty workplace → text chapters → 10s ref2v clips → one mute-free join** No storyboard grids. No contact sheets. No TTS, no `audio_url`, no generated BGM. The lead stays a **cat body**. Gags are **acted**. Speech is **off**—clip foley stays on the join. Costume stills are identity plates, not opening keyframes. ## What you get | For creators | For the production stack | | --- | --- | | **Beginner-first** — no photo? the agent invents a job-character still from the idea you clicked and runs toward the cut | **Opinionated pipeline** — language lock, identity-plate gate, ref2v packing, and unmute join encoded in-skill | | **Cat-first** — your upload is the face; no photo is a generate path, not a dead end | **ref2v-first** — no accidental i2v / t2v / fl2v / director / first-frame chains | | **Job that reads on sight** — apron, badge, desk, paws on the tool; jokes are actions, not captions | **Two-ref binding** — `input_urls` = character plate, then empty workplace; neither still is the first or last frame | | **Dynamic ideas** — look + job + gag buttons written for this turn, not a canned trio | **Short prompts** — `[reference generation]` first, then a dense English clip body | | **Delivery that continues** — after clip 1, buttons to finish / change a beat / redo; after the cut, next moves in your language | **Progressive disclosure** — craft lives in `references/`, not bloated context | | **Plain creator voice** — short lines about this cat and this job; no pipeline jargon unless you ask | **Marketplace-ready** — clear triggers, defaults, and tool contracts | ## Pipeline | Stage | Output | | --- | --- | | **Intake** | Lock language + **9:16** + default **30s** (3×10s). Photo → bind that cat. No photo + idea chosen → **one `t2i`** job character (look + costume). Vague ask → clickable ideas, then generate | | **Identity** | Character still as a **plain-background plate** (not sitting at the desk). Upload path: `i2i` costume onto the photo only | | **Workplace** | Empty set still (`t2i`, `locations`) — no cat in the plate | | **Text chapters** | N × 10s cards (default 3) — identity, job, gag, screenshot hero pose. No grids | | **ref2v** | One 10s clip per chapter. Prompt starts with `[reference generation]`. Character + workplace as **references**, not keyframes | | **Join** | `mix_clips` once, `volume=1`. **No mute. No music tool. No BGM overlay.** 10s-only skips the join | ## Production stack | Parameter | Default | Notes | | --- | --- | --- | | **Episode** | **30 seconds** | 3 clips × 10s. Optional 10s (1 clip) or 20s (2 clips) if you ask | | **Engine** | `ref2v` · `reference_to_video` | Multi-ref identity lock. Forbidden: `i2v` / `ai2v` / `t2v` / `fl2v` / `director` | | **Clip duration** | **10 seconds** | Always set. Chapters are separate clips, then one join | | **Video orientation** | Portrait · **9:16** | Landscape only if you ask. Character still, workplace still, and video share the lock | | **Image refs** | Character plate always; empty workplace always | Character = `t2i` `characters` (no photo) or `i2i` from upload. Workplace = `t2i` `locations` | | **Ref order** | Character → workplace | Identity plate, not a first-frame. Forbidden: single-image calls; empty desk as opening frame; frozen replica of either still | | **Identity plate** | **On, every no-upload / costume still** | Portrait, costume, **plain studio background**. Forbidden: cat already at the desk in the still | | **Cat lock** | **Cat body** | Apron / badge / small hat OK. Forbidden: human body, fur-color drift, swapping cats after lock | | **Speaker lock** | **N/A — no speech** | No dialogue, no VO, no `generate_tts`, no burned-in captions | | **Dialogue** | **Off** | Gags are visible verbs (paws, props, desk), not subtitle jokes | | **Picture audio** | **Clip foley, kept** | Light workplace foley from the video. Join does **not** mute. No `generate_music` | | **Cut grammar** | **3×10s then join** | `mix_clips` `volume=1`. Forbidden: mute join; second pass overlay BGM | | **BGM** | **Off** | No music tool. Do not ask the video model to invent a score | | **Confirm** | **After clip 1** | Default pause so you can confirm the cat. One-shot / generate-all skips that pause | | **Buttons** | **Written this turn** | Not a fixed catalog. Ideas, jobs, and next moves named from this cat and this gag | | **Language** | **User-typed lock** | Replies, buttons, todos, asset names in your language; image/video prompts stay English | ## Modes | Mode | You bring | Agent delivers | | --- | --- | --- | | **Full auto-run** | “Make a cat workday vlog” / generate all | Invent a look + job if needed → stills → 3 clips → unmute join | | **Idea-first** | One-line premise (“orange tabby night-shift coder”) | One `t2i` job character → empty desk → chapters → cut | | **Photo-first** | Your cat photo | Bind that face; `i2i` light costume only; never replace the cat | | **Pick-an-idea** | Nothing locked yet | 2–4 concrete cat+job+gag buttons in your language; click → generate that character | | **Job swap** | “Make it a barista instead” | New costume / workplace; same locked cat | | **Single-clip** | “Just 10 seconds” | One ref2v clip; skip the join | | **Single-stage edit** | “Redo clip 2” / “Change the desk” / “New gag” | That stage only, then remake the affected clip / cut | | **One-shot** | “Finish in one go / no confirm” | No pause after clip 1; still pause on the finished cut | | **Lengthen** | “20 seconds” / “30 seconds” | 2 or 3 × 10s chapters, then one unmute join | ## Example prompts **Full auto-run** > Make a cat workday vlog. Portrait, 30 seconds. **Idea-first** > Orange tabby night-shift coder. Paws slam the keyboard, then it falls asleep on the keys. Vertical. **Photo-first** > This is my cat. Make a cute office comedy — barista, 9:16. **Pick-an-idea** > I don’t have a photo. Give me a few cat-and-job ideas and just start. **Job swap** > Same cat. Change the job to convenience-store night shift. **One-shot** > Generate everything in one go. Don’t stop after the first clip. **Single-stage edit** > Redo the middle clip — the gag should be the foam overflowing, not the keyboard. ## When to use - Vertical cute workplace shorts: one cat at a job, contrast-cute, acted gags - You uploaded a cat photo, **or** you only typed an idea - You want to change the job, rewrite shots, redo one clip, or remake the cut - Creators who want a finished 30-second join with original clip sound, not a mute beauty reel > **Not the right tool when** you need a plain pet portrait, a cat turned into a human, a product ad, a spoken / lip-sync track, a generated song overlay, storyboard grids for client sign-off, or real animal harm. ## Invoke ``` /cat-workday-vlog ``` Or: *"Turn this into a 30-second portrait cat workday — same cat, job that reads on sight, ref2v, keep the clip sound, no mute mix."* **Your cat · Your job · One coherent workday.**
ShortDrama
109 0 4
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/3325120/2026-09-18/acyduyne.jpg
--- name: cat-workday-vlog display_name: Cat Workday Vlog description: Use when the user wants a vertical cute workplace comedy short about a cat at a job (cat office vlog, job gags, contrast-cute anthropomorphic cat), with or without an uploaded cat photo; also use to change the job, rewrite shots, or redo one clip. allowed-tools: - recognize_image - generate_image - generate_video - mix_clips - create_document - update_document canvas: hidden_entity_types: - work_in_progress --- # Cat Workday Vlog Make a **portrait 9:16 workplace comedy short** starring one cat at a job. Prefer the user's cat photo when they uploaded one. If they did not, invent the job character with one text-to-image still, then continue. Costume, props, and set must make the job obvious. Gags are acted, not subtitled. End on a screenshot-worthy workday pose. **Aspect (hard):** default `orientation='portrait'` (9:16) on every still and clip. Do not ask. Landscape only if the user names it. This Skill has **no** structured project state. Keep facts in the reply or `create_document`. Reuse canvas URLs. **Video channel (hard):** every clip uses `generate_video(model='ref2v', gen_mode='reference_to_video')`. Forbidden: `i2v` / `t2v` / `ai2v` / `fl2v` / `director` / physical model names. `gen_mode='reference_to_video'` must be passed with `ref2v`. `input_urls` = **job-character still, then empty workplace** (identity + set references). Never put an audio URL in `input_urls`. This is **reference-to-video**. The prompt **must start with** `[reference generation]`. Forbidden: treating any input still as the first/last frame, `[image to video]`, first-frame / keyframe completion. **Zero CJK in this package.** Skill files stay English (handbook only). User-facing runtime copy follows `user_lang`. Image and video `prompt` fields stay English. Image and video prompts follow packing in `references/02-assets.md` and `references/04-video.md`. Keep those prompts short English. ## Global Language Lock (mandatory, highest priority) Follow the **user's typed sentences**, not platform chrome. **Chrome trap:** strip the slash-invoke line, `<untrusted_skill_content>`, `[session_assets]`, tool traces, and injected English such as "The synthesis failed…". **CJK wrapper ≠ `zh-CN`. English Skill dump ≠ `en`.** Count Han only in leftover `typed_user`. Zero Han → not Chinese (English words → `en`) and **zero CJK** in every user-visible string. Majority Han → `zh-CN` / `zh-TW`. Translate user copy at runtime — never paste handbook English, and **never put Han characters into this Skill**. **Two layers:** `user_lang` (everything the user can read) and MODEL English (`generate_image` / `generate_video` `prompt`). Algorithm: `references/00-language.md`. On the **first user message**, silently `read_skill_file` **only** `references/00-language.md` once, then lock `user_lang` **before any user-visible sentence and before any `generate_*` / `create_document`**. Do not re-detect after tools, English files, `[session_assets]`, or "continue". Do **not** also load later stage files on that turn. The first user-visible sentence must already be in `user_lang` and about **this cat / this job / this vlog**. | Layer | Scope | Language | |---|---|---| | **User-facing** | Replies, buttons, todos, asset `name`, canvas body | `user_lang` | | **Image / video prompt** | `generate_image` / `generate_video` `prompt` | **English always** | `zh-CN` vs `zh-TW` is a lock. One turn = one `user_lang`. If a mix is found, rewrite before send. Platform English never re-detects. Switch `user_lang` only when the user explicitly asks to change reply language. ## When to use - The user wants a cat workday vlog, office-cat gag, or anthropomorphic pet job short - The user uploaded a cat photo, **or** only typed an idea with no photo - The user only wants to change the job, redo one clip, or remake the cut Do not use for: plain pet portraits, turning the cat into a human, product ads, or real animal ha