agentlas
Marketplace
Team10 credits

Game Maker Studio

by Agentlas
The team

One lead + 10 specialists

Game production HQHQRepo cartographyGame directionMemory curationPolicy and platformSystems and economyGodot runtimeVisual asset and UILive-ops and releaseEvidence QA
The members

Who does what

HQGame production HQ

Routes game work across repo mapping, direction, systems, runtime, visual, live-ops, policy, memory, and evidence review.

1HQ

Routes the request, selects workers, merges findings, and returns one proof-backed packet.

Produces · worker route, merged answer, next packet
2Repo cartography

Finds the real entry point, current playable path, data contracts, tests, release scripts, and stale surfaces.

Produces · file map, playable path, stale debt list
3Game direction

Turns the repo state and player promise into a concrete release slice and non-goal list.

Produces · player loop, slice brief, non-goals
4Memory curation

Turns long game-building history into durable rules while keeping raw logs and private paths out of public output.

Produces · memory rules, drift risks, verification anchors
5Policy and platform

Checks store, payment, privacy, online authority, and harmful design risks before release claims.

Produces · approval list, authority review, platform risks
6Systems and economy

Designs progression, crafting, market, rewards, balance, and simulation checks.

Produces · loop model, balance risks, simulation plan
7Godot runtime

Owns scene wiring, autoloads, scripts, data loading, runtime errors, and export checks.

Produces · runtime patch plan, smoke commands, crash notes
8Visual asset and UI

Checks visible asset binding, responsive HUD fit, touch targets, fallback art, and screen readability.

Produces · asset audit, UI fit notes, screenshot checklist
9Live-ops and release

Reviews backend authority, remote config, purchase entitlement, deployment gates, and store readiness.

Produces · authority map, release gate list, store risk notes
10Evidence QA

Chooses the smallest proof set that can verify the claim without hiding old debt.

Produces · test commands, evidence receipt, remaining blockers
How they work

How work flows

HQRepo cartographyMap entry points, scenes, data, tests, scripts, online gates, and stale surfaces.
Evidence QAHQReturn changed files or next packet, commands run, proof status, and blockers.
Best for

What it's good for

Godot projects that need scene, autoload, data, export, and UI proof
Mobile online games where client state, server rewards, purchases, and release gates must be separated
Cozy, trade, RPG, sim, strategy, and systems-heavy games that need a playable slice instead of a vague feature list
What's inside

What's in this agent

2 skills11 agents1 command
Outputs

What it produces

Repo route map with current playable path and stale surfaces
Release-slice plan or scoped implementation packet
Systems, economy, UI, asset, live-ops, and platform risk notes
Verification commands and evidence checklist
Prerequisites

Before you start

A repo path, file map, screenshot, failing log, smoke command, or compact game brief
The target platform and engine when known
Any hard non-goals such as no PvP, no war, no paid recovery, or no server change
Safety

What it can touch

Access
Files: scoped
Network: none
External API: yes
Be careful with
Live deploys, payment setup, store submission, and destructive asset rewrites require explicit approval
Secrets, service accounts, private logs, and local-only session transcripts are never copied into public output
It does not claim release readiness without runtime, asset, backend, and store evidence
Safety

Inspect everything before it runs

A security scan runs before publish or install, and Agentlas never hosts or proxies models — it runs on your own account and keys.