New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Eva Launch Video

by vvedantb in vvedantb/eva

Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL…

MITAuto-check passedMedia & Creative

Install Eva Launch Video

skills CLI
$ npx skills add vvedantb/eva --skill eva-launch-video -a claude-code

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

GitHub CLI
$ gh skill install vvedantb/eva eva-launch-video --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/vvedantb/eva.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/eva-launch-video .claude/skills/eva-launch-video && 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
eva-launch-video
GitHub stars
101
Token cost
~2.7k tokens
SKILL.md length
1,276 words
Files
3 (incl. references)
Skills in repo
23
Repo updated
First seen
Licence
MIT

At a glance

Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL…

  • Works in 7 steps: 1280×720, 30fps. Not 1080p. The eva UI… → Capture at the frame's native layout,… → Real footage, finished states. Drive the… → …
  • Re-render an eva demo/product/launch/Twitter video
  • SKILL.md covers Non-negotiable defaults…, Workflow, Quick reference and Existing compositions, plus 1 more section
  • Calls node, npx and pnpm

What it does

Eva Launch Video is an agent skill from vvedantb/eva. Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL running app via agent-browser (no mockups). Use this whenever the user wants to make, update, re-cut, re-score, re-capture, or re-render an eva demo/product/launch/Twitter video or screen-recording walkthrough, or mentions the video/ Remotion project — even if they don't say "Remotion". It encodes hard-won defaults…

Its SKILL.md is about 2.7k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files, including reference files (for example `references/capture-workflow.md` and `references/remotion-build.md`).

It sits in Media & Creative, covering Video production, Browser automation and UI design. It works with Remotion and X (Twitter). The repository describes itself as: Orchestrate sandboxed agents that run in the cloud while you work. The licence is MIT.

When your agent uses it

  • Re-render an eva demo/product/launch/Twitter video
  • Screen-recording walkthrough
  • Mentions the video/ Remotion project — even if they dont say Remotion

Example prompts

  • “t say”
  • “/eva-launch-video”

Requirements

  • Node.js

Workflow steps

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

  1. 1280×720, 30fps. Not 1080p. The eva UI is a dense desktop app; at 1080p it reads "too zoomed /
  2. Capture at the frame's native layout, show full-bleed. Set the browser viewport to 1280×720
  3. Real footage, finished states. Drive the actual app with agent-browser. Record
  4. Snappy, hard-cut, beat-synced. ~20s total. Cut on the beat (use , not long
  5. Music synced to every cut. The user's standing preference is **lo-fi, the "classic" variant, no
  6. Match eva's theme. Near-black #050606, indigo accent #818cf8, brand mark gradient purple
  7. Be honest + protect confidentiality. Use the user's OWN repo (vvedantb/eva), never client

What it can do on your machine

Read from SKILL.md and the folder at commit a5a4df2. 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

    Shell commands in SKILL.md call:

    • node
    • npx
    • pnpm
    • npm
    • bash

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

  • Network

    No URLs in SKILL.md. Its commands use npx, pnpm and npm, which can reach the network depending on how they are called.

    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

Eva Launch Video loads about 2.7k tokens when it runs, and up to ~6k if it reads all its reference files. Until then it costs about 159 tokens; SKILL.md has 1,276 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~159
When it runs · the whole SKILL.md, loaded when a task matches
~2.7k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~6k

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); files beside SKILL.md are not scanned.

SKILL.md

The full file from vvedantb/eva at commit a5a4df2, republished under its MIT licence (© vvedantb). 1,276 words, ~2,692 tokens.

Download SKILL.mdSave it as .claude/skills/eva-launch-video/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
eva-launch-video
description
Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL running app via agent-browser (no mockups). Use this whenever the user wants to make, update, re-cut, re-score, re-capture, or re-render an eva demo/product/launch/Twitter video or screen-recording walkthrough, or mentions the `video/` Remotion project — even if they don't say "Remotion". It encodes hard-won defaults (resolution, zoom/framing, theme, pacing, music sync, rendering, QA) so you don't repeat past mistakes.

eva product video

Make a high-end product video of the eva app (Linear/Notion quality) using the existing Remotion project at video/. Footage comes from the real app, captured with agent-browser.

Wrong skill? This one is the polished multi-scene edit (Remotion, music, captions). For a single feature, prefer the lighter siblings: eva-feature-screenshot (one screenshot + tweet) or eva-feature-demo (one raw agent-browser recording + tweet). Tweet copy for any of the three comes from eva-tweet.

The canonical project already exists — reuse it, don't start over.

  • Project: video/ · primary composition: EvaHero (video/src/EvaHero.tsx)
  • Render output: video/out/eva-demo.mp4
  • Reusable components in video/src/components/: Backdrop, AppWindow, EvaMark, HeroCaption (+ Accent), Cursor, Spotlight, StatCounter, SnapIn, tokens.
  • Helper scripts in video/scripts/: stills.mjs, gen-music.mjs, check-audio.mjs.

To make a NEW video: re-capture screens for the new story, edit the scene list / captions / durations in EvaHero.tsx, regenerate the music to match the cut times, render, QA.

Non-negotiable defaults (learned the hard way — change only with a reason)

Each of these was a mistake first. Keep them unless the user overrides.

  1. 1280×720, 30fps. Not 1080p. The eva UI is a dense desktop app; at 1080p it reads "too zoomed / too big" on a phone. 720p is the mobile-friendly sweet spot for Twitter/X.

  2. Capture at the frame's native layout, show full-bleed. Set the browser viewport to 1280×720 when capturing, then display the screenshot full-bleed at 1:1 (filling the whole frame). The single biggest quality lever:

    • Capturing at 2560×1440 and downscaling to 720p halves UI text size → "zoomed out" and unreadable on mobile.
    • Cropping/zooming the camera into a region shows awkward fragments of the UI.
    • A 1280-wide layout shown 1:1 = native element sizes = readable + "zoomed in" + crisp.

    Zero zoom means zero. PanScene from/to stay at { scale: 1.0, fx: 0.5, fy: 0.5 } for every screenshot scene. Even a 1.02–1.06 "subtle push-in" clips the chrome the viewer needs — the sidebar labels, the primary button in the top-right corner, the stats footer. Also pass <SnapIn from={0.97}> around screenshots: the default 1.06 pops inward and crops the opening frames of every cut. Motion comes from the hard cuts and caption snaps, not from camera moves.

  3. Real footage, finished states. Drive the actual app with agent-browser. Record already-completed work (merged tasks, diffs, PRs) — do NOT wait on live agent runs (sandboxes are stopped; live runs are slow and flaky). Stage clean inputs (e.g. a tidy task title in the New Task modal). See references/capture-workflow.md.

    Never submit — type only. Fill the input, open the picker, capture. No Enter in a composer, no send arrow, Create, Save, Run, or Start, and never kick off a real session or agent run, not even for demo footage. Submitting mutates the user's real workspace and burns real compute; the staged state is all the shot needs. Enter is allowed only where it commits to the field itself and nothing else (accepting an @-mention from an open picker) — otherwise click the picker row.

  4. Snappy, hard-cut, beat-synced. ~20s total. Cut on the beat (use <Series>, not long crossfades) and snap captions in fast. Pop scenes in with SnapIn — bare over title/outro cards, from={0.97} over screenshots (see default 2). Boring = too long + too many slow crossfades.

  5. Music synced to every cut. The user's standing preference is lo-fi, the "classic" variant, no clap — generate it with gen-lofi-variants.mjs, which marks each cut with a warm-chord swell instead. It takes --duration, --cuts, --only, --prefix, so a new video needs no edits to the script:

    bash
    node scripts/gen-lofi-variants.mjs --duration=20.5 --cuts=2.5,5.5,8.5,11.5,14.5,17.5 \
  --only=classic --prefix=platform      # → public/audio/platform-music-lofi-classic.wav

    --cuts must match the scene boundaries in the composition. gen-music.mjs is the older clap bed — only use it if the user asks for one. See references/remotion-build.md → Music.

  6. Match eva's theme. Near-black #050606, indigo accent #818cf8, brand mark gradient purple #8B3FB8 → blue #3B7DD8, fonts Inter + JetBrains Mono. All in video/src/components/tokens.ts. Never invent brand colors.

  7. Be honest + protect confidentiality. Use the user's OWN repo (vvedantb/eva), never client repos (Evalucom/CarePulse). Don't show the "Codebases" home (lists client repos), and check document/task lists for client-named items before capturing — stage a clean demo doc if the list is dirty. Never raise personal-data concerns about names or emails on screen — Settings → Experimental has a blur-PID toggle that blurs every [data-pii] element, so this is handled in the product. Do flag error panels leaking env var names or project ids. Captions describe truthfully — never imply a feature the screen doesn't show. If the only available capture is a stopped sandbox or an empty state, reword the caption to match the screen rather than dressing it up.

Show full SKILL.md (568 more words)Show less

Workflow

Work in this order; read each step's reference file before doing the step.

  1. Plan the story. Default hero flow (8 scenes, ~20s): title → describe a task → Eva writes code → review & ship PR → board → projects → proof (count-up stats) → outro. Pick the scenes and one-line captions first.

  2. Capture footage → read references/capture-workflow.md. Start the app (pnpm dev), sign in as the agent user, set viewport 1280×720, screenshot each screen into video/public/captures/.

  3. Compose / edit → read references/remotion-build.md. Edit EvaHero.tsx: scene list, durations (multiples of the beat), captions, the cursor path for the input scene. Reuse existing components.

  4. Score it → read references/remotion-build.md → Music. Run gen-lofi-variants.mjs with --cuts matching the scene boundaries (default 5 above), point the composition's <Audio> at the generated file, confirm MUSIC = true.

  5. Render + QA. You can't preview the mp4 directly (no system ffmpeg, and you can't watch video) — so QA with still frames, one per scene, taken mid-scene:

    bash
    cd video
node scripts/stills.mjs "40,110,200,290,380,470,560" EvaPlatform   # 2nd arg = composition id

    The composition id defaults to EvaHero — pass it explicitly or you will QA the wrong video. Read the PNGs in video/out/ and check every edge of the frame: is the top-right primary button whole, are the sidebar labels whole, does the caption overlap live UI text? Then render + verify:

    bash
    npx remotion render EvaPlatform out/eva-platform.mp4 --codec=h264 --image-format=png
node scripts/check-audio.mjs out/eva-platform.mp4   # duration, 1280×720, audio track present

    --image-format=png keeps UI text crisp. Iterate on stills until it's right, then render.

Quick reference

NeedCommand (run from video/)
Live preview / scrubnpm run dev (Remotion Studio)
Capture one screenbash scripts/shot.sh /vvedantb/eva/web/sessions sessions-list 4000
Render a compositionnpx remotion render <Id> out/<name>.mp4 --codec=h264 --image-format=png
QA specific framesnode scripts/stills.mjs "40,110,200" <CompositionId>
Regenerate musicnode scripts/gen-lofi-variants.mjs --cuts=… --only=classic --prefix=<name>
Verify output filenode scripts/check-audio.mjs out/<name>.mp4

Existing compositions

EvaHero (hero flow) · SessionsDemo (cloud sessions) · EvaPlatform (all five surfaces in one place: sessions → docs → testing arena → kanban tasks → PR reviews) · QuickTasksDemo (legacy). Register any new one in video/src/Root.tsx at 1280×720 / 30fps.

Gotchas that cost time before

  • Sign-in URL is /?agent=true (boolean), not /?agent — the route validates agent as a zod boolean and /?agent throws "Expected boolean, received string". If the callback hangs on "Signing in…" with You're already signed in, navigate straight to the target page.
  • Dev overlays land in the screenshot. react-scan renders an FPS toolbar into a shadow root on #react-scan-root, and agentation into [data-agentation-root]. A stylesheet alone doesn't kill react-scan — its host carries inline styles, so also set display:none !important inline on the host. scripts/shot.sh does both; use it rather than hand-rolling a capture chain.
  • Chain agent-browser calls in one Bash command with timeout, and background anything long. A naive open → wait → screenshot chain blows the tool timeout. Refs from snapshot -i go stale the moment the dev toolbar expands — re-snapshot after hiding overlays, before clicking.
  • Pin all @remotion/* packages to the same exact version or the bundler errors.
  • Music can't be auditioned by Claude. The bed does exactly what's asked (a swell on each cut) but tell the user to listen and swap for a licensed track if they want — set MUSIC=false for silent.
  • Sandboxes in old session captures are stopped, so every session screen reads "Start the sandbox to…". Don't caption it as live. Nothing in captures/sessions/ shows a running sandbox.
  • Captions carry the message — most Twitter playback is muted, so every scene needs a caption.

Detailed how-tos: references/capture-workflow.md (footage) and references/remotion-build.md (composition, theme, music, rendering, QA).

© vvedantb, 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 2 other files (references) in .claude/skills/eva-launch-video of vvedantb/eva.

  • SKILL.md
  • references/capture-workflow.md
  • references/remotion-build.md

Open the folder on GitHubat commit a5a4df2

Compare with similar skills

Eva Launch Video 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.

Eva Launch Video compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Eva Launch Video this skillvvedantb/eva101—~2.7kAutomated safety check: PassMIT
Convertremotion-dev/remotion63k—~247Automated safety check: PassCustom licence
Docsremotion-dev/remotion63k—~248Automated safety check: PassCustom licence
Player Exampleremotion-dev/remotion63k—~258Automated safety check: PassCustom licence
Studioremotion-dev/remotion63k—~376Automated safety check: PassCustom licence
Video Craftsendaifun/solana-new178—~1.5kAutomated safety check: PassMIT

Similar skills

  • Convert

    remotion-dev/remotion

    Official

    Start the local @remotion/convert app and open it in the Codex browser.

    63k GitHub stars~247 tokensUpdated today
    Media & CreativeAuto-check passed

  • Docs

    remotion-dev/remotion

    Official

    Start the Remotion docs site and open it in the Codex browser.

    63k GitHub stars~248 tokensUpdated today
    Media & CreativeAuto-check passed

  • Player Example

    remotion-dev/remotion

    Official

    Start the @remotion/player example app and open it in the Codex browser.

    63k GitHub stars~258 tokensUpdated today
    Media & CreativeAuto-check passed

  • Studio

    remotion-dev/remotion

    Official

    Start Remotion Studio from packages/example and open it in the Codex browser.

    63k GitHub stars~376 tokensUpdated today
    Media & CreativeAuto-check passed

  • Video Craft

    sendaifun/solana-new

    Frame-level visual composition and product demo presentation for Remotion videos.

    178 GitHub stars~1.5k tokensUpdated 7 days ago
    Media & CreativeAuto-check passed

  • Cy Release Video

    compozy/compozy

    Create or revise CompozyOS release videos with high-resolution product captures, Ultramock camera scenes, and the reusable Remotion package.

    2.8k GitHub stars~1.2k tokensUpdated yesterday
    Media & CreativeAuto-check passed

More from vvedantb/eva

All 23 skills in this repo

  • Eva Feature Demo

    vvedantb/eva

    Record a real agent-browser screencast of one eva feature being used end to end, convert it to an X-ready mp4, and write a tweet for it.

    101 GitHub stars~3.3k tokensUpdated today
    Auto-check passed

  • Animate

    vvedantb/eva

    Build an animation from scratch, making the decisions in the order that determines whether it feels right — should it animate at all, what purpose, which tool, which properties, which curve and…

    101 GitHub starsUsed in 6 repos~2.9k tokens
    Auto-check passed

  • Convex Create Component

    vvedantb/eva

    Design and build Convex components with clear boundaries, isolated state, and app-facing wrappers.

    101 GitHub stars~3.3k tokensUpdated today
    Auto-check passed

  • Eva Feature Screenshot

    vvedantb/eva

    Grab a single clean HD screenshot of a new eva feature from the real running app (Playwright at deviceScaleFactor 2, 1280 layout captured crisp at 2560×1440, dev overlays hidden) and write a tweet…

    101 GitHub stars~3.2k tokensUpdated today
    Auto-check: notes

  • Code Structure

    vvedantb/eva

    A skill your agent uses when multiple workflows duplicate the same operational logic, when deciding what belongs in actions vs shared services, or when refactoring repeated operational blocks across…

    101 GitHub starsUsed in 2 repos~1.1k tokens
    Auto-check passed

  • Libraries Dev

    vvedantb/eva

    Use the Libraries.dev UI effect libraries correctly and find where they fit in a project.

    101 GitHub stars~2.5k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Eva Launch Video

What does Eva Launch Video do?

Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL…. Eva Launch Video is an agent skill from vvedantb/eva. Produce polished, mobile-friendly product demo videos of the eva app with Remotion — 1280×720, snappy beat-synced hard cuts, lo-fi music that swells on every cut, and footage captured from the REAL running app via agent-browser (no mockups).

When should I use Eva Launch Video?

Eva Launch Video fits situations like: re-render an eva demo/product/launch/Twitter video; screen-recording walkthrough; mentions the video/ Remotion project — even if they dont say Remotion.

How do I install Eva Launch Video in Claude Code?

Run `npx skills add vvedantb/eva --skill eva-launch-video -a claude-code`. Or copy the skill folder (.claude/skills/eva-launch-video in vvedantb/eva) into .claude/skills/eva-launch-video in your project. Claude Code loads it when a task matches its description.

How do I install Eva Launch Video in Codex?

Run `npx skills add vvedantb/eva --skill eva-launch-video -a codex`. Or copy the skill folder (.claude/skills/eva-launch-video in vvedantb/eva) into .agents/skills/eva-launch-video in your project. Codex loads it when a task matches its description.

Can I use Eva Launch Video 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 vvedantb/eva --skill eva-launch-video -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/eva-launch-video, .gemini/skills/eva-launch-video, .github/skills/eva-launch-video and .opencode/skills/eva-launch-video in your project.

What does Eva Launch Video need to run?

Going by SKILL.md and its folder, Eva Launch Video needs the command-line tools its instructions call (node, npx, pnpm, npm and bash). Our summary lists: Node.js.

Does Eva Launch Video access the network?

SKILL.md contains no URLs. Its commands use npx and npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Eva Launch Video 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. Review the folder before installing.

What licence does Eva Launch Video use?

Eva Launch Video 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 Eva Launch Video use?

About 2.7k tokens (SKILL.md is roughly 11k 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 3.3k tokens, read only when the agent opens those files.

What are the alternatives to Eva Launch Video?

Skills that share tags, products or a category with Eva Launch Video: Convert (remotion-dev/remotion, 63k stars), Docs (remotion-dev/remotion, 63k stars), Player Example (remotion-dev/remotion, 63k stars) and Studio (remotion-dev/remotion, 63k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Eva Launch Video?

vvedantb (a GitHub user) maintains it in vvedantb/eva, which has 101 GitHub stars. The repository holds 23 skills in this directory. The repository was last updated on October 9, 2026.

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