anabologyco-maker/claudia-bot/mv-studio
Claude Code and agent skill by anabologyco-maker/claudia-bot
- community
Checking that the SKILL.md file is still in the repository.
Use when the owner wants a Studio-tier video (their own Suno song and Midjourney stills, Seedance sung clips with measured lip sync, 30-45 turns, $50-110), when a project's tier is studio, or to find out which stage, skill and gate come next.
Install
npx vskill@latest install anabologyco-maker/claudia-bot/mv-studioSKILL.md
Studio tier - the stage machine
Read docs/00-overview.md §2, §4 and §5 once per project. Stages run in order, each stage has a skill, and the owner speaks only at the gates. Needs an OpenRouter key or the Higgsfield connector (sung clips); neither: offer skill mv-quick.
Every new conversation: slop --json -p NAME status, then slop --json -p NAME decide --list --last 20, then continue at next. After each gate, suggest a new chat for the next stage.
Plan files (once per project)
- PRODUCTION.md: copy the skeleton in docs/examples/production-skeleton.md into the project folder; fill the go message verbatim, the rules, the budget per pool and the stage table (docs/95-delegation.md §6). Write a change of plan there first.
- DECISIONS.md: every owner message that decides something, verbatim and dated:
slop --json -p NAME decide '<their words>' --by owner. A changed decision: add--supersedes '<the old one>'; never edit old entries.
Stages
| # | Stage | Skill | Gate, recorded with | |
|---|---|---|---|---|
| 1 | brief + taste capture | taste | - | |
| 2 | concept, 3 pitches | song | G1 concept: decide | |
| 3 | lyrics, Suno pack, rounds by ear | song | G2 song: decide | |
| 4 | ingest, analysis, ear check | song | - | |
| 5 | lead canon + 4 test plates | plates | G3 face: decide | |
| 6 | storyboard | storyboard | G4: board approve | |
| 7 | Midjourney book, paste, ingest, picks | plates | G5: gate prompts | |
| 8 | clip prompt book in one context | clips | G5 again for the clips | |
| 9 | clips, sync, covers | clips | - | |
| 10 | edit, render, scan, v1, master, 9:16 | render-and-review | G6: gate v1 |
Rules (docs/00-overview.md §4 and §5.3)
- Lock the song (G2) and finish the ear check before the storyboard. Stage 5 can run beside stages 2-4.
- Spend: show the book and the estimate (generated seconds x price x 2); wait for an explicit "go". Silence never approves spend. One action over the confirm limit needs its own yes.
- Combine gates that are ready into one message (the face sheet with the Suno pack or a song round). The board may pass on silence only if the owner set that window themselves; record it.
- Before the full v1 show a 20-40 s excerpt (skill render-and-review step 6) to catch "too gray" early; the cheaper look first:
slop --json -p NAME job start -- render sheet --chapter 1 --review. - A quality complaint:
slop --json -p NAME spend --freeze '<their words>'before anything else. - Sung clips: screenplay for acting and motion, short for copy: every sung window takes the short sung form (docs/60-clips.md §4.0). Sung plates show the whole mouth, nothing in front of it (no microphone, hand or hair). Takes from a long prompt that did not copy: retake short on the same provider before any re-route. Photoreal close-ups on Higgsfield, stylized or wide on OpenRouter (§6.5). Never alter a declined image.
- Long runs: wait by the profile's turn rule, and record them:
slop --json -p NAME runs add ID --what '<what>' --resume '<command>'so a later conversation can resume. - Ping the owner between gates only for a failing paid service, a handoff (login, captcha, download) or a decision that is theirs.
Too many turns for the owner's plan, or EV-class bespoke graphics wanted: offer skill mv-director.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/mv-studio/SKILL.md
Popularity
- Repository stars
- 73
- Repository forks
- 14
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/anabologyco-maker/claudia-bot/mv-studio)