Agent skill

Stage Compose

by Orkas-AI in Orkas-AI/Orkas-VideoStudio

Authoring knowledge for Orkas/OVS HTML video compositions -- write an index.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an…

MITAuto-check passedMedia & Creative

Install Stage Compose

skills CLI
$ npx skills add Orkas-AI/Orkas-VideoStudio --skill stage-compose -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install Orkas-AI/Orkas-VideoStudio stage-compose --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/Orkas-AI/Orkas-VideoStudio.git skills-src && mkdir -p .claude/skills && cp -r skills-src/packages/skills/stage-compose .claude/skills/stage-compose && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
stage-compose
GitHub stars
498
Token cost
~4.2k tokens
SKILL.md length
2,171 words
Files
6 (incl. scripts, references)
Skills in repo
14
Repo updated
First seen
Licence
MIT

At a glance

Authoring knowledge for Orkas/OVS HTML video compositions -- write an index.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an…

  • Works in 9 steps: Read the approved… → If standalone narration is needed,… → Keep the approved… → …
  • Explainer / animation / motion-graphics / caption / lower-third / title-card work
  • SKILL.md covers Fast COMPOSE Runbook, HTML Preview Gate, Check And Repair Policy and Render (The Outcome), plus 3 more sections
  • Runs JavaScript scripts from its folder

What it does

Stage Compose is an agent skill from Orkas-AI/Orkas-VideoStudio. Authoring knowledge for Orkas/OVS HTML video compositions -- write an index.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an mp4. Trigger for explainer / animation / motion-graphics / caption / lower-third / title-card work, or to build a compose segment inside an AUTO plan; do NOT trigger for cutting real footage (stage-edit) or AI footage (stage-generate).

Its SKILL.md is about 4.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 7 other files, including scripts and reference files (for example `references/art-direction.md`, `references/manifest-and-authoring.md` and `references/narration.md`).

It sits in Media & Creative, covering Motion graphics and AI video generation. It works with HeyGen. The repository describes itself as: Turn your coding agent into a video studio: describe a video in plain language, and your agent writes the timeline and produces the file. The licence is MIT.

When your agent uses it

  • Explainer / animation / motion-graphics / caption / lower-third / title-card work
  • Build a compose segment inside an AUTO plan
  • Do NOT trigger for cutting real footage (stage-edit)
  • AI footage (stage-generate)

Example prompts

  • “/stage-compose”

Requirements

  • Node.js

Workflow steps

9 steps, taken from the first numbered list in SKILL.md.

  1. Read the approved project/composition/composition-manifest.json and this skill if not already loaded for the current turn. Also read…
  2. If standalone narration is needed, capture the exact ovs speech-capabilities profile in the approved plan and run ovs narration fit before…
  3. Keep the approved project/composition/composition-manifest.json with schema_version: 2, including approved timeline/copy/source mappings…
  4. Decide whether to open the optional HTML Preview Gate before rendering mp4. Use the preview gate when expected render rework is expensive…
  5. If visual preview is needed, run ovs check and ovs snapshot. Review the complete contact-sheet index and open the cover, QA-named frames…
  6. Run the draft command: ovs draft project/composition --out project/render/draft.mp4 --quality draft --report…
  7. If draft fails, repair the highest canonical source, reconcile when needed, and retry only after the authored signature changes. Do not…
  8. If the draft command returns ok: true, the composition is frozen for final-video confirmation. Do not edit index.html…
  9. If preview was skipped and the fallback trigger applies, run composition-design-review against representative draft frames. Open…

What it can do on your machine

Read from SKILL.md and the folder at commit c0c3c3e. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (JavaScript), which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Stage Compose loads about 4.2k tokens when it runs, and up to ~7.8k if it reads all its reference files. Until then it costs about 112 tokens; SKILL.md has 2,171 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~112
When it runs · the whole SKILL.md, loaded when a task matches
~4.2k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~7.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from Orkas-AI/Orkas-VideoStudio at commit c0c3c3e, republished under its MIT licence (© Orkas-AI). 2,171 words, ~4,179 tokens.

Download SKILL.mdSave it as .claude/skills/stage-compose/SKILL.md (or your agent's skills folder). This skill also uses 5 other files; get the full folder from GitHub.
name
stage-compose
description
Authoring knowledge for Orkas/OVS HTML video compositions -- write an index.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an mp4. Trigger for explainer / animation / motion-graphics / caption / lower-third / title-card work, or to build a compose segment inside an AUTO plan; do NOT trigger for cutting real footage (stage-edit) or AI footage (stage-generate).

stage-compose

How to author an Orkas/OVS HTML composition and turn it into a video. composition-manifest.json v2 is the canonical timeline/audio/design artifact. In this open-source build, ovs draft is the VideoStudio-style production gate: it runs manifest/source/narration/local-asset checks, HyperFrames check/render, media QA, sampled-frame video QA, and writes one report. HyperFrames remains the render backend; do not bypass the draft gate for user-facing drafts.

For visual direction, apply frontend-design before writing manifest art_direction. If the user provides a reference image/video, DESIGN.md, brand guide, screenshot, design notes, existing app UI, or named style, apply design-system-importer to convert it into intent-bound reference constraints and compact tokens. composition-design-review is a bounded full-frame review before a visual preview is shown, with a post-draft fallback when preview is skipped.

Gate authorization belongs to gate-control. This skill supplies the canonical manifest, contact sheet, draft, and QA evidence; after any Gate B/Preview/Gate D decision or resumed approval, run ovs gate transition rather than inventing a second confirmation or recovery loop.

When the production runtime is unavailable, return a clearly unexecuted candidate package instead of fabricating a render: locked assumptions, final-form script/narration, timed storyboard, exact visible copy/captions, visual/audio and rights-safe asset plan, target export settings, and checkable preview/final QA.

Fast COMPOSE Runbook

After Gate B approves the canonical composition manifest, keep the production turn narrow:

  1. Read the approved project/composition/composition-manifest.json and this skill if not already loaded for the current turn. Also read frontend-design; read design-system-importer only when a concrete style source or explicit named reference exists. Read composition-design-review after snapshot evidence exists, or after the draft when preview was intentionally skipped and the fallback trigger applies.
  2. If standalone narration is needed, capture the exact ovs speech-capabilities profile in the approved plan and run ovs narration fit before synthesis. Visual authoring, ovs check, and ovs snapshot may proceed while narration is pending so the visual candidate remains reviewable. Required narration blocks only complete ovs draft/final delivery. Before drafting, run ovs speak once to project/composition/assets/narration.mp3, declare the composition-owned track, probe its duration, rerun measured fit, and reconcile. Reuse an existing matching narration file; do not synthesize again just because a visual preview changed. For an AUTO segment, use audio.owner: "assembler" and render silent.
  3. Keep the approved project/composition/composition-manifest.json with schema_version: 2, including approved timeline/copy/source mappings, audio ownership, and art_direction. Confirm its dedicated cover, any references/reference_fidelity, and VisualDirectionV1; then prepare and author static resolved/hero frames before deterministic motion. The exact 0s cover must already be readable. Reconcile after manifest timing/audio changes without replacing authored DOM/CSS/SVG/motion.
  4. Decide whether to open the optional HTML Preview Gate before rendering mp4. Use the preview gate when expected render rework is expensive: target duration >= 20s, scene count >= 3, render cost is likely slow, or the composition has dense text, complex SVG/GSAP, many branded/supplied assets, tight narration timing, or a prior draft failure. Skip it for short/simple work: target duration < 20s, scene count <= 2, no narration/timing complexity, and no obvious visual-risk signal. The subject category alone never forces the preview gate.
  5. If visual preview is needed, run ovs check and ovs snapshot. Review the complete contact-sheet index and open the cover, QA-named frames and risky cells at full size; compare declared references. Collect all blockers, apply one localized repair, then rerun check/snapshot and review the complete new revision. The contact sheet is only an index. The author checklist is advisory and adds no scored approval gate; present current passing snapshot evidence after addressing concrete visible blockers.
  6. Run the draft command: ovs draft project/composition --out project/render/draft.mp4 --quality draft --report project/render/draft-report.json --findings project/composition/qa/check.json. Before rendering, this gate validates manifest/HTML consistency, prepares declared local vendor assets, blocks remote runtime resources, verifies local assets, checks shotlist/source alignment, and checks narration mapping. Then it runs HyperFrames check/render, media QA, sampled-frame QA, and writes one report.
  7. If draft fails, repair the highest canonical source, reconcile when needed, and retry only after the authored signature changes. Do not delete QA state or repeat an unchanged strategy. After two non-converging passes, preserve the current artifact and show the visible unresolved issue with concrete options through gate-control; await a real user direction before another cycle.
  8. If the draft command returns ok: true, the composition is frozen for final-video confirmation. Do not edit index.html, composition-manifest.json, assets, or narration again in the same turn unless the report contains a real blocker (check/source/audio/video QA failure), or the user explicitly asks for a revision. Visual/readability warnings and design-review fix/polish notes are advisories, not permission to self-repair.
  9. If preview was skipped and the fallback trigger applies, run composition-design-review against representative draft frames. Open final-video confirmation only after the draft command returns ok: true and any triggered review has no concrete blockers.

The default path is model-authored HTML -> draft. Do not write or compile spec.json; fixed template compilation is not part of the COMPOSE path because visual quality and extensibility come first.

HTML Preview Gate

Use the HTML Preview Gate to avoid expensive mp4 rerenders when visual rework is likely. It is a cost-control gate, not a new creative milestone, and it is only for the COMPOSE line. Decide from expected rework cost:

  • Preview first when duration >= 20s or scene count >= 3.
  • Preview first when a shorter piece has dense text, multiple chapters, many supplied/brand assets, complex SVG/GSAP motion, tight narration timing, or a prior draft/repair failure.
  • Skip preview when duration < 20s, scene count <= 2, and the HTML is simple enough that rendering the draft is cheaper than asking for another confirmation.
  • Do not use product/promo/version-update labels alone as the trigger. Those labels only contribute to risk when the piece is long, visually dense, or expensive to rerender.

When preview is triggered, run ovs check and ovs snapshot, review the complete contact-sheet index and open the cover and risky frame_paths with composition-design-review, and show only a fully reviewed revision. The contact sheet is an index; full-size frames resolve doubtful cells. Show the full frame evidence, the index.html path, and a compact status line:

  • reason for preview: duration / scene count / complexity / prior failure
  • check headline: blocking count or main advisory
  • what approval means: render mp4 draft next

If the user revises, edit the manifest or HTML, reconcile when needed, then rerun check/snapshot and the complete frame review. Keep this loop lightweight; reuse materialized narration and do not render mp4 during visual preview.

The HTML Preview Gate does not replace the mp4 draft. It cannot validate audio muxing, final encoded video quality, sampled-frame video QA, or exact narration pacing. After approval, always run ovs draft and open Gate D with the video.

Read manifest and authoring before writing or reconciling manifest/HTML. It owns the canonical schema, offline assets, semantic hooks and HyperFrames timing contract.

Read art direction before authoring the visual system. It owns the concrete design budgets and reference constraints.

Read render and preview before judging reference-led snapshot or draft evidence. It owns paired source/target comparison and the boundary between contact-sheet indexing and full-size inspection.

Check And Repair Policy

Run the draft command before any user-facing video. If structural/source/audio/video QA fails, repair the highest canonical source and retry only after its signature changes. After two non-converging passes, preserve the evidence and request a concrete user direction through gate-control. Technical errors with repair passes remaining do not create confirmation requests.

Repairs should address the cause, not just the symptom:

  • FONT_TOO_SMALL: reduce text density, shorten copy, enlarge/reflow containers, or move labels out of small shapes. Do not simply increase every font size if that creates overflow.
  • missing_timeline_registry, gsap_timeline_not_registered: register a paused GSAP timeline on window.__timelines[compositionId], using the exact root data-composition-id.
  • timed_element_missing_clip_class, root_composition_missing_data_start, media_missing_data_start, imperative_media_control: let the renderer own timing and media playback through data-start, data-duration, .clip, and media data attributes.
  • text_occluded, text_box_overflow, content_overlap: restructure the scene layout or regenerate the affected scene from the contract's boxes. Do not rely on small numeric nudges.
  • FROZEN_FRAME_RUN: treat repeated sampled frames as an advisory until semantic evidence confirms the timeline is frozen. Inspect the visual evidence; repair timeline registration, scene clip timing, or variation only when the frames actually fail to change as intended.

If only visual advisories remain and the draft render exists, present the mp4 draft with QA notes instead of silently looping. Repair the manifest or hand-authored HTML directly; do not introduce spec.json as a workaround.

After the draft render succeeds, use composition-design-review only as the fallback when preview was skipped and its trigger applies. A review blocker must be visible in a specific scene/frame and must break readability, the approved promise, required brand/style tokens, motion timing, or asset safety. Allow at most one localized repair and re-draft only for concrete blockers. Treat fix and polish findings as final-video confirmation notes.

Read narration and audio only when this composition carries speech or other audio. It owns narration fitting, existing-speech preservation and audio ownership.

Show full SKILL.md (718 more words)Show less

Render (The Outcome)

Produce the finished video by running the draft gate over the composition directory. Iterate at draft quality, then do one high-quality pass once the layout and timing pass review:

  • Draft: ovs draft project/composition --out project/render/draft.mp4 --quality draft --report project/render/draft-report.json --findings project/composition/qa/check.json
  • Final: ovs draft project/composition --out project/render/video.mp4 --quality high --report project/render/final-report.json --findings project/composition/qa/final-check.json

Use ovs render only as a narrow diagnostic render when QA has already identified the blocker. User-facing drafts and finals go through ovs draft so contract, source, narration, media, and sampled-frame QA are not skipped.

After final-video approval, a local visual-only revision reuses the approved plan, assets, and narration. Change only the affected scene, run reconcile when needed, then ovs check, ovs snapshot, full-frame design review, and one revised high-quality encode. Do not request production-plan confirmation again or synthesize narration/generation again unless the requested edit changes signed copy, timing, language, narration, source mapping, or provider intent.

Director Judgment (Compose Line)

Craft calls specific to designed/animated explainers, on top of the shared craft reference (video-craft):

  • One concept per visual chapter -- do not stack two ideas in one scene; give each its own build.
  • Concrete before abstract -- real data, diagrams, steps before a metaphor; the metaphor only lands once the concrete version is understood.
  • Aesthetic thesis before styling -- use frontend-design to choose one signature visual device that comes from the subject matter; spend distinctiveness there and keep the rest disciplined.
  • Reference intent governs fidelity -- use design-system-importer for DESIGN.md/brand/reference input. For reproduce/edit, preserve declared geometry, visual relationships, and authorized assets except named may_change; for guide, adapt only the declared roles. Respect explicit exclusions without forcing unrelated variation.
  • Design review is a pre-preview guardrail -- inspect the entire snapshot frame set before showing it; when preview is skipped, use the triggered post-draft fallback. Block only on concrete visible failures; template feel, hierarchy, and polish issues that do not break the promise become final-video confirmation notes.
  • Render exact text as real text -- stats, names, CTAs are typed into the composition, never baked into AI imagery.
  • Build to the narration words, not arbitrary beats; hold a fully-built scene/chart >= 2-3 s before moving on.
  • Vary scene types -- no three near-identical layouts in a row; alternate full-frame / split / diagram / quote.
  • Spoken/readable captions live in the plan's tracks.captions.lines (data), NOT burned into this composition -- the assembler burns them via burnsubs at the end, so a later typo fix is a one-line edit, not a re-render of the whole composition. Only a purely decorative caption treatment that is the visual design may live inside the composition.
  • Surface craft-threshold warnings on the composition when you QA it before rendering. In draft mode, small readable text is a QA advisory rather than a render blocker; oversized palette is advisory and should be judged against the design thesis, brand, and scene clarity.

Constraints

  • Deterministic only: no real-time timers, no network-dependent runtime behavior, no randomness without a fixed seed -- the renderer seeks discrete frames.
  • Keep all referenced assets inside the composition directory so the render is self-contained.
  • This skill authors and renders compositions; it does not pick the production line (see video-router) or generate AI footage.

Scene timing and narration identity

Position authored tweens with scaffold S("scene-id") and D("scene-id"), which read canonical scene attributes after reconcile. Never move HyperFrames-owned clip visibility into custom playback code. The delivered opening must already carry visible title/hero content at exactly zero; animate from a visible base, not a blank fade. Every child also starts with visible content; only the delivered opening needs the whole-video cover promise.

An explicit empty narration_text marks a silent scene and clears stale narration references. Preserve those silent windows when fitting speech. Shorter speech that fits the existing spoken windows leaves the visual timeline unchanged. Keep the declared target separate from measured audio; do not truncate speech or silently change a fixed user duration. Narration-only changes reuse visual approval only if all visual assets and scene windows remain identical. New pixels or timing require a current complete preview.

Reuse speech only when text, route, voice, model, format, speed and output bytes match its receipt. Keep one file per line in EDIT/AUTO. Their signed start_sec and target_sec are fixed windows: shortening text or using an approved retry fits the audio into the window; moving the timeline requires the corresponding user authority. Music/SFX ownership alone does not imply narration.

© Orkas-AI, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 5 other files (scripts, references) in packages/skills/stage-compose of Orkas-AI/Orkas-VideoStudio.

  • SKILL.md
  • references/art-direction.md
  • references/manifest-and-authoring.md
  • references/narration.md
  • references/render-and-preview.md
  • scripts/composition.mjs

Open the folder on GitHubat commit c0c3c3e

Compare with similar skills

Stage Compose next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Stage Compose compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Stage Compose this skillOrkas-AI/Orkas-VideoStudio498—~4.2kAutomated safety check: PassMIT
HyperFrames Video Entry Pointheygen-com/hyperframes59k3 repos~5.2kAutomated safety check: PassApache-2.0
Hyperframes ShowreelBlackBeltTechnology/pi-agent-dashboard315—~1.5kAutomated safety check: PassMIT
Super Video MakerBomx/super-video-maker-skill309—~11kAutomated safety check: NotesNone
AI Motion DirectorPluviobyte/video-production-skills667—~1.6kAutomated safety check: PassNone
Avatar MixUpload-Post/avatar-mix131—~4.9kAutomated safety check: NotesMIT

Similar skills

  • HyperFrames Video Entry Point

    heygen-com/hyperframes

    Entry point for making, editing and rendering videos from HTML compositions with HyperFrames, routing each request to the right workflow.

    59k GitHub starsUsed in 3 repos~5.2k tokens
    Media & CreativeAuto-check passed
  • Hyperframes Showreel

    BlackBeltTechnology/pi-agent-dashboard

    Turn real screen and site footage into a rhythmic showreel rendered with HyperFrames: sampling, edit script, redacted clips, a generated composition, Veo only as additive FX, music edit + camera…

    315 GitHub stars~1.5k tokensUpdated today
    Media & CreativeAuto-check passed
  • Super Video Maker

    Bomx/super-video-maker-skill

    End-to-end AI video production skill for agentic frameworks.

    309 GitHub stars~11k tokensUpdated 2 mo ago
    Media & CreativeAuto-check: notes
  • AI Motion Director

    Pluviobyte/video-production-skills

    Motion-first AI video director for topic-to-video creation. An agent skill from Pluviobyte/video-production-skills.

    667 GitHub stars~1.6k tokensUpdated 2 mo ago
    Media & CreativeAuto-check passed
  • Avatar Mix

    Upload-Post/avatar-mix

    Crea un video 16:9 con tu avatar HeyGen (tu cara + tu voz) presentando contenido, con montaje dinamico que alterna entre avatar a pantalla completa, avatar en esquina sobre un fondo, y solo fondo…

    131 GitHub stars~4.9k tokensUpdated 2 mo ago
    Media & CreativeAuto-check: notes
  • Collage B-Roll Video Maker

    imraywang/rayskills

    Turns a spoken line or a full topic into editorial-collage videos of halftone cut-outs assembled piece by piece, as 5-second B-roll or a narrated explainer.

    159 GitHub stars~2.6k tokensUpdated 16 days ago
    Media & CreativeAuto-check passed

More from Orkas-AI/Orkas-VideoStudio

All 14 skills in this repo
  • Gate Control

    Orkas-AI/Orkas-VideoStudio

    Canonical VideoStudio review authorization and state-transition policy.

    498 GitHub stars~2.7k tokensUpdated 17 days ago
    Auto-check passed
  • Stage Edit

    Orkas-AI/Orkas-VideoStudio

    Intelligent editing of real user-supplied footage—understand it with transcript/inspected-frame/scene/silence/quality evidence, then choose deterministic timeline operations or a constrained…

    498 GitHub stars~2.4k tokensUpdated 17 days ago
    Auto-check passed
  • Stage Consistency

    Orkas-AI/Orkas-VideoStudio

    Multi-shot narrative & character consistency — a character bible with a locked front-portrait anchor, view-matched reference selection, recent-frame carry-forward, Cameo (a user photo as the lead)…

    498 GitHub stars~1.8k tokensUpdated 17 days ago
    Auto-check passed
  • Stage Plan

    Orkas-AI/Orkas-VideoStudio

    The "ingest + plan" half of end-to-end video orchestration — ingest the user's material from evidence, then decompose intent into ONE cross-modal EDL (plan.json: edit/generate/compose/provided…

    498 GitHub stars~4k tokensUpdated 17 days ago
    Auto-check passed
  • Frontend Design

    Orkas-AI/Orkas-VideoStudio

    Aesthetic direction for OrkasVideoStudio HTML and motion-graphics compositions.

    498 GitHub stars~4.3k tokensUpdated 17 days ago
    Auto-check passed
  • Video Craft

    Orkas-AI/Orkas-VideoStudio

    The craft standard that makes a video GOOD, not just rendered — hooks, story pacing, visual hierarchy, type & safe zones, motion/easing, captions, audio mix, platform conventions, shot language…

    498 GitHub stars~3.3k tokensUpdated 17 days ago
    Auto-check passed

Works with

Questions about Stage Compose

What does Stage Compose do?

Authoring knowledge for Orkas/OVS HTML video compositions -- write an index.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an…. Stage Compose is an agent skill from Orkas-AI/Orkas-VideoStudio.html, drive animation from a paused timeline, declare canvas + duration, then run the VideoStudio draft gate to render an mp4.

When should I use Stage Compose?

Stage Compose fits situations like: explainer / animation / motion-graphics / caption / lower-third / title-card work; build a compose segment inside an AUTO plan; do NOT trigger for cutting real footage (stage-edit); AI footage (stage-generate).

How do I install Stage Compose in Claude Code?

Run `npx skills add Orkas-AI/Orkas-VideoStudio --skill stage-compose -a claude-code`. Or copy the skill folder (packages/skills/stage-compose in Orkas-AI/Orkas-VideoStudio) into .claude/skills/stage-compose in your project. Claude Code loads it when a task matches its description.

How do I install Stage Compose in Codex?

Run `npx skills add Orkas-AI/Orkas-VideoStudio --skill stage-compose -a codex`. Or copy the skill folder (packages/skills/stage-compose in Orkas-AI/Orkas-VideoStudio) into .agents/skills/stage-compose in your project. Codex loads it when a task matches its description.

Can I use Stage Compose in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Orkas-AI/Orkas-VideoStudio --skill stage-compose -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/stage-compose, .gemini/skills/stage-compose, .github/skills/stage-compose and .opencode/skills/stage-compose in your project.

What does Stage Compose need to run?

Going by SKILL.md and its folder, Stage Compose needs JavaScript for the scripts in its folder. Our summary lists: Node.js.

Does Stage Compose access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Stage Compose safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Stage Compose use?

Stage Compose is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Stage Compose use?

About 4.2k tokens (SKILL.md is roughly 17k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3.7k tokens, read only when the agent opens those files.

What are the alternatives to Stage Compose?

Skills that share tags, products or a category with Stage Compose: HyperFrames Video Entry Point (heygen-com/hyperframes, 59k stars), Hyperframes Showreel (BlackBeltTechnology/pi-agent-dashboard, 315 stars), Super Video Maker (Bomx/super-video-maker-skill, 309 stars) and AI Motion Director (Pluviobyte/video-production-skills, 667 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Stage Compose?

Orkas-AI (a GitHub user) maintains it in Orkas-AI/Orkas-VideoStudio, which has 498 GitHub stars. The repository holds 14 skills in this directory. The repository was last updated on September 22, 2026.

Source: Orkas-AI/Orkas-VideoStudio on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.