Stickman Animation Ad Skill for AI Agents

The whole thing

One prompt in. A finished stickman film out.

Nothing below was edited by hand. Three ten-second clips, each landing on the next one's opening frame, with the narration, the score and the overlays the skill wrote itself.

Make a stickman explainer for Hevy. Light theme, 16:9, about 30 seconds.

Start creating Install the skill

Your brand, your script, the same director.

Just ask

Say the word, it runs.

The skill fires on natural requests like these. No commands to memorize.

“Make a stickman ad for my brand”“Turn this script into a stick-figure explainer”“Create a kinetic line-animation video”“Make a doodle-style comic video”

The pipeline

What the skill does for you.

1
Pick the theme and format
The skill reads the brand and asks for the theme (white canvas with black figures, or black canvas with white figures) and the aspect ratio. The brand color becomes one of up to three accent colors, each with a single meaning.
2
Approve the director's proposal
Narration, then one row per 10-second clip: three timed beats, at least four visual devices, camera moves, sound effects, and the exact state each clip hands to the next. Nothing renders until you approve it.
3
Draw the opening frames
One flat line-art frame per clip, anchored on the first so line weight and proportions hold. Each frame is both the start of its clip and the landing point of the clip before it.
4
Animate clip to clip
Every clip is animated from its own frame and lands on the next one, so transitions happen in-camera. Prompts hold the flat look, forbid on-screen text, and keep clips free of voices.
5
Narrate, score, and compose
One narrator reads each clip's line on its visuals, music follows the emotional turn, overlays are added as crisp text, and an optional on-twos snap finishes the limited-animation feel.

Three beats in every ten seconds.

Each clip establishes, escalates and hands off, so nothing stands still.

Start creating

Light or dark

Two themes. One strict look.

The skill asks before it draws: white canvas with black figures, or black canvas with white. Then it locks the rest. Faceless figures, one line weight, a flat canvas, and at most three accent colors, each with one job.

Theme and ratio are asked at kickoff, never picked silently
Your brand color becomes an accent and owns the product and the payoff
No words inside any frame; on-screen phrases are crisp overlays added at the end

Start creating Install the skill

AllTrails

Strava

Blue Tokai

The model stack

The right model for each job.

GPT Image 2 draws each clip's flat line-art frame by default, and Gemini Omni Flash 1.1 animates it, landing every clip on the next clip's frame. Seedance 2.0 takes the rare move past 10 seconds. Ask for GPT Image 2.5 Sunburst or Seedance 2.5 and the agent uses it for every clip. The voice and the score are made separately, so the narrator never changes between clips.

What you get

A finished narrated stick-figure video in 9:16, 16:9, or 1:1
A director's proposal with timed beats and named transitions
One opening frame and one animated clip per 10-second section
Separate narration, music, and an editable composition

Start creating Read the source

GPT Image 2.5 SunburstNew

On request, crisper line art for every opening frame

GPT Image 2

The flat line-art opening frame for every clip

Nano Banana 2

The backup for a single frame that keeps picking up shading

Gemini Omni Flash 1.1New

Each 10-second clip, from its own frame to the next one

Seedance 2.5New

On request, one unbroken move up to 30 seconds

Seedance 2.0

The rare move that has to run past 10 seconds

xAI TTS

The narrator, one line per clip in the same voice

MiniMax 2.6

The music, which turns when the story does

Install

Add it in seconds.

Install the skill, then make sure the Advibly MCP is connected so it has tools to call. Works with every skills-capable agent.

Claude Code

Claude

Cursor

Codex

VS Code

OpenClaw

Hermes

ChatGPT

Connect the Advibly MCP first

One click, no terminal

Download .skill Download .zip

Drop the .skill file into Claude → Settings → Skills, or unzip the folder into ~/.claude/skills.

One command (any agent)

npx skills add advibly/skills -s stickman-animation

Works with Claude Code, Claude Desktop, Cursor, and any agent that supports skills. Pulls straight from GitHub.

Manual (Claude Code)

git clone https://github.com/advibly/skills.git
cp -R skills/stickman-animation ~/.claude/skills/

Or in Claude Desktop: Settings → Skills → Install and select the skill folder.

Bundled references

The playbook.

Each skill ships with reference docs the agent loads on demand, so the craft is baked in.

director-proposal.md

Story patterns, the three-beat visual-density recipe, transition interfaces, composition by aspect ratio, and the proposal format.

storyboard-prompts.md

Light and dark style locks, the opening-frame scaffold, the flat product prop, and frame QA.

animate-prompts.md

The clip contract for Gemini Omni Flash 1.1: frame hold, timed beats, landing state, and negatives within 2000 characters.

audio-and-gotchas.md

Per-clip narration, music, the composition call with overlays and cadence, and a symptom-to-fix table.

example.md

One worked example: proposal excerpt, an opening frame, two clip prompts, and the composition call.

Honest limits

What to watch for.

The visual system is intentionally strict: a flat monochrome canvas, faceless line-art figures, and at most three accent colors.
Generated frames and clips carry no words; any on-screen phrase is added as an overlay in the composition.
End-frame landings are close but not frame-exact, so a seam can need a small trim in the editor.
Narration is generated separately; video clips stay voice-free so the narrator never changes between clips.

Questions

Before you install it.

What does a video from this skill look like?

Faceless line-art stick figures on a flat white or black canvas, with up to three saturated accent colors and a visual change every two to three seconds. Each 10-second clip is built as three beats: establish, escalate, then hand off to the next clip, which picks up exactly where it ended.

What do I give the agent?

A brand set up in Advibly, plus a product, an angle, or a script. It asks for the theme (light or dark), the ratio (9:16, 16:9 or 1:1) and, if it isn't obvious, the length: about 30 seconds is three clips, about 60 is six. Then it writes a director's proposal and renders nothing until you approve it.

How long does a full video take?

The gates are the slow part. You approve the director's proposal, then the opening frames. After that the clips render in parallel, the narration and music are generated, and the finished video lands in your Advibly library with a composition you can reopen in the editor.

What does it cost?

The skill is free and open source. Running it spends your normal Advibly credits on the frames, clips and audio it generates, with no surcharge for the skill, and every generation shows its credit cost before it runs. The proposal includes a credit estimate too.

Which agents can run it?

Any agent that supports installable skills and the Model Context Protocol: Claude Code, Claude, Cursor, Codex, VS Code, OpenClaw, Hermes and ChatGPT. Install with one npx command, then connect the Advibly MCP so the skill has tools to call.

Can it put words on screen?

Yes, but never inside the animation. Video models garble text, so no frame or clip carries words, numbers or labels. Any on-screen phrase is added as a crisp text overlay in the final composition. The one exception is your product's own logo, copied from its photo.