Stellar Anomaly - Between Projects

How a small team, zero supervision, and questionable sleep created a retro space survival shooter

Every studio has that strange, dangerous gap between projects. Not long enough to relax. Not short enough to ignore. Just enough time for a dev team to stare at each other and collectively decide to make a terrible decision.

Ours was simple: let’s build a full game with whatever spare time we can steal. No roadmap meetings. No leadership steering. No sacred document full of “requirements” that kills the soul in chapter one.

We gave the crew free reign, a rough theme, and permission to get weird. The result was inevitable.

Stellar Anomaly was born. A mission-based deep space survival shooter with progression, bosses, biomes, endless mode, and a story that dips its toes into quantum mechanics—because nothing says “healthy hobby project” like arguing about reality at 1:37 AM.

We set rules that look cute on paper and become violent in practice:

• Short spare time only. Tiny windows. No heroic all-nighters. (We lied.)
• Needs to be done fast. Fast enough that doubt doesn’t have time to move in.
• Minimal assets. Most visuals should come from code. PNGs only when absolutely necessary.
• Free or AI-generated assets only. If it isn’t free, it doesn’t exist.

It wasn’t just “make a game.” It was prove we can ship a complete experience while deliberately limiting the usual escape hatches: content bloat, asset packs, and that classic move where you “temporarily” use a placeholder and forget it for six months.

Quantum mechanics is one of those topics that drags you into a bar fight with the universe. The more you learn, the more you realize reality has the audacity to behave like a prank.

We’re not quantum physicists. We’re builders with curiosity and enough ego to poke the void with a stick. So we built a story that treats quantum weirdness the same way we treat production: respect it, fear it, and ship anyway.

The premise is simple: a corrupted star core called Helios Sigma is bending the rules. Space stops behaving. Biomes mutate. Bosses don’t just hit hard—they feel like they’re cheating.

And that’s the fun part. You’re not just dodging bullets. You’re navigating a universe that can’t decide what it is today.

This is where it gets spicy. We didn’t run this like a typical project. We ran it like a survival simulation of teamwork.

No leadership input. No constant guidance. People just took ownership of parts they cared about. Ideas got tossed in. The best ones survived. The bad ones got executed behind the station. Quietly.

The constraint was the manager. The timeline was the manager. The fact we wanted it to be finished was the manager.

Minimal assets changes your brain chemistry. Suddenly you’re not “finding art.” You’re manufacturing atmosphere:

• Animated starfields, storms, haze, distortion—built with particles and shaders-by-willpower
• Biome identity created through palettes, motion, and hazard behavior
• UI polish through glassy gradients, neon edges, and clean animation timing
• Boss presence through scale, pattern design, phase transitions, and screen effects

And when we did use assets, we treated it like contraband: only free assets or AI-generated, only when necessary, and always with purpose.

Generating assets with AI is easy. Generating the right assets is an exercise in learning how to communicate like a machine that’s allergic to context.

We learned fast: prompt engineering isn’t “typing magic words.” It’s writing requirements the way you’d brief a contractor you’ve never met, who may or may not hallucinate a third moon into your logo.

# Copy, paste, fill in. This is how you stop random surprises.

TITLE: "Stellar Anomaly - [ASSET NAME]"

STYLE: "Retro sci-fi, clean silhouettes, high contrast, neon accents, minimal noise"
FORMAT: "PNG with transparent background, crisp edges, no text"
PALETTE: "Teal (#00ffd5), emerald (#00c588), magenta (#ff006e), deep navy (#000814)"

SUBJECT:
"- Single object centered
- Must read clearly at 128px height
- Simple geometry, avoid tiny details"

DO:
"- Use a bold silhouette
- Keep lighting subtle
- Make it feel 'game UI friendly'"

DON'T:
"- No busy backgrounds
- No photorealism
- No extra props
- No random symbols or letters"

OUTPUT CHECK:
"If it won't work as a sprite, regenerate."

That template saved us hours. More importantly, it prevented the dreaded situation where someone says, “Yeah, we’ll just fix it later,” and later never arrives.

Under the hood, it’s a full arcade experience with campaign structure, progression systems, and enough content to keep your mouse hand angry for a while.

We built it for desktop, optimized for mouse precision, and kept it browser-based because we like the idea that you can just launch it and instantly start making bad life choices in space.

Here’s the uncomfortable truth: small ad-hoc projects don’t fail because of tech. They fail because people don’t feel ownership. Or they feel too much ownership and start a religious war over folder naming.

We got lucky—and also smart. Everyone contributed. People grabbed what they wanted: a biome, a weapon, a system, a UI pass, a sound pass, mission logic, boss patterns, progression flow. Nobody waited for permission.

And because there were no strict requirements, there was no hiding. You couldn’t “work on something” forever. You either shipped it into the build, or it stayed as a beautiful thought.

That pressure creates clarity. Not the panic kind. The focused kind. The “we’re doing this now” kind.

• You can adapt to a short timeline and still deliver something complete
• You can build most of the “wow” with code and smart design choices
• You can use free assets and AI assets without your game looking like a ransom note
• You can get strong results when teammates have freedom and responsibility
• You can make a story about reality breaking without breaking your pipeline