12 Free BRUTAL Framework Checks Before You Fork - 2026

If you searched brutal framework checks after 15 October 2026, you do not need another “open source engine” trailer. You need a numbered free-check kit before anyone forks ahwootech/BRUTAL over a live game — or treats a star click as a migration.

Ahwoo and RocketWerkz open-sourced BRUTAL as a C# foundation: NuGet packages, native interop including Vulkan, no royalties, and a license that still wants credit. The README and Ahwoo BRUTAL page already name .NET 10, Vulkan 1.3, Windows or Linux getting-started, and the hard not-list (no editor, no scene graph, no renderer). Missing is a Discord-pasteable pass that forces those checks before a fork becomes the production story.

This listicle owns the free kit: twelve checks that cost nothing but time (middleware like FMOD may still have their own fees — the checks are free). It is not the what-changed map, not the Unity replacement Stop, and not a license-only credit lab. Those own facts, refusal, and fill-in evenings. This URL owns the toolkit roundup.

Why this matters now

Studios fork because Waylith and Kitten Space Agency exist, then discover no scene view, no renderer, a logo splash they never budgeted, or a Unity CI job they deleted on vibes. The README already listed the pieces. What people skip is a minimum gate before the fork is treated as the live stack.

Direct answer: Run checks 1–4 before you promise Discord anything. Run 5–9 on a side repo with Hello + one owned spike. Run 10–12 before you tell partners the shipping title “moved.” If check 3 fails (not-list unread), stop — that is the point of the kit.

Who this list is for

Time: 25 minutes for 1–4; 2–4 hours for a full lab pass.
Honest limit: README, LICENSE.txt, and Ahwoo pages are authorities. Package floors move — re-check NuGet the morning you adopt.

Non-repetition

Pair a Discord paste with 25 free game development Discord servers.

Four-check minimum (if you only have twenty-five minutes)

Do not fork “over” a shipping game until these four are green.

The twelve checks

1. .NET 10 SDK literacy

Do: Install or verify the .NET 10 SDK. Run dotnet --version. Match the README getting-started era.

Pass: Version noted in the lab doc with date.
Fail: “We have some .NET.”
Do next: If fail, stop — fix SDK before NuGet shopping.

2. README not-list latch

Do: Open the README section What BRUTAL is not. Confirm aloud: no editor, no scene graph, no renderer, PRs that drift will be rejected.

Pass: Ticket quotes the three nos.
Fail: Roadmap still says “Prefabs next sprint via upstream.”
Do next: Feed the Unity Stop if anyone still says replacement.

3. License 2.1–2.3 sticky

Do: Read BRUTAL License v1.0. Write three stickies: credits copyright (2.1), logo before gameplay (2.2), ship NOTICE.txt (2.3). Bookmark Ahwoo branding for the logo asset.

Pass: Stickies or doc section signed by compliance owner.
Fail: “It’s MIT, we’re fine.”
Do next: Orient on the BRUTAL guide.

4. Shipping-engine KEEP

Do: Write UNITY_KEEP=true (or Godot/Unreal KEEP) for the live title. Fork and prototype only in a side repo name.

Pass: Standup can see the KEEP latch.
Fail: Plan deletes the only green CI job.
Do next: If pressure remains, open the replacement Stop.

5. OS + GPU prereq smoke

Do: Confirm lab OS matches README getting-started (Windows or Linux emphasized). If rendering, confirm a Vulkan 1.3 capable GPU and current drivers. Optional: LunarG Vulkan SDK for validation layers.

Pass: OS + GPU notes in lab doc.
Fail: Promising three-desktop parity from a Mac-only anecdote without checking package targets.
Do next: Re-read ahwoo.com/brutal target list; verify before pitch decks.

6. NuGet restore on a clean machine / CI agent

Do: dotnet new console, add Ahwoo.Brutal.Core (and later Glfw). dotnet restore must succeed on a machine that is not your daily driver — preferably CI.

Pass: Green restore log attached to the ticket.
Fail: “Works on my laptop with cached packages.”
Do next: Document corporate proxy / TLS breaks; package restores fail the same way across stacks.

7. Hello window (Core + Glfw)

Do: Follow the README first window sample: init GLFW, 1280×720, poll until close, terminate. No graphics API required for this check.

Pass: Screenshot of the empty window + commit hash.
Fail: Calling stars a vertical slice.
Do next: Only then schedule a render spike.

8. Minimal package intent (not kitchen sink)

Do: Write the next three packages you will add and why (e.g. Vulkan abstractions, ImGui, Concurrency). Do not reference every package in the README table on day one.

Pass: Intent list in the lab doc.
Fail: Solution with twenty packages and no owned triangle.
Do next: Add one cluster per spike.

9. Owned spike beyond Hello

Do: Clear color, one triangle, or one texture path you understand. Log which packages and which code you own vs upstream.

Pass: Short write-up + artifact.
Fail: “We’ll use Waylith numbers as our estimate.”
Do next: If spike fails, KEEP the shipping engine without drama.

10. Third-party license scan

Do: For any audio/net/physics package you plan (FMOD, etc.), open that package’s LICENSES notes. Section 4.0 of BRUTAL’s license says third-party terms still apply.

Pass: List of extra licenses in the diligence folder.
Fail: Assuming BRUTAL’s royalty-free line covers FMOD commercial terms.
Do next: Budget middleware correctly.

11. Comms redlines

Do: Draft Discord/trailer lines that match reality: side prototype, not “we left Unity.” Ban “drop-in Unity alternative” and “no credit needed.”

Pass: Paste approved by producer.
Fail: Store page rewrite on announcement day.
Do next: Use the Stop’s redline list as a checklist.

12. Fork-over-live gate

Do: Before any fork is treated as production: G1 systems-you-own list, named render + tools owners, license plan signed, shipping CI still green, producer sign-off. Until then PRODUCTION_FLAG_BRUTAL=false.

Pass: Signed gate sheet.
Fail: “Fork is the product now.”
Do next: Remain on side-repo literacy.

Beginner evening path

1. Checks 1–4 on paper (30–40 min).
2. Check 6–7 Hello window (60–120 min).
3. Check 3 branding bookmark + NOTICE path draft (20 min).
4. Paste Discord brief; leave live engine unchanged.
5. Schedule check 9 on another evening.

Common mistakes

• Forking the monorepo into the Steam build folder.
• Skipping 2.2 because “open source.”
• Kitchen-sink NuGet on day one.
• Measuring success only by GitHub stars (2037 by 20 Oct 2026 in our evidence pass — attention, not readiness).

Working-dev pipeline order

Discord paste (creators)

BRUTAL — 12 free checks before you fork:
1 .NET 10  2 README not-list  3 license 2.1–2.3  4 UNITY_KEEP
5 OS/GPU  6 NuGet restore  7 Hello window  8 package intent
9 owned spike  10 third-party licenses  11 comms redlines
12 fork-over-live gate
Min: 1–4. Kit: /blog/12-free-brutal-framework-checks-before-you-fork-2026

Receipt before treating a fork as production

Date:
.NET 10 version:
Not-list quoted?: Y/N
2.1/2.2/2.3 plan:
Shipping KEEP?: Y/N
OS/GPU notes:
Restore CI log:
Hello commit:
Next three packages:
Spike artifact:
Third-party license list:
Comms approved?: Y/N
PRODUCTION_FLAG_BRUTAL?: Y/N
Signer:

Failure modes

Free vs not free

RACI

Worked example — wrong fork Friday

Setup: Engineer forks BRUTAL into the game repo, deletes Unity CI, posts “we’re on BRUTAL.” Art asks for the scene view. Legal has never seen 2.2.

What the kit forces: 1–4 before promises; 6–7 before pride; 3+11 before trailers; 12 before production language.

Solo vs studio

Solo: Four-check minimum same night; Hello on a second folder; never rename your shipping repo to BRUTAL on socials.

Studio: Producer owns check 12; do not let a star spam override KEEP.

Sibling map

When to re-run the twelve

Re-run 1 when SDK floors bump. Re-run 3 before every public BRUTAL-linked build. Re-run 6 after CI image changes. Re-run 11 before every trailer. Re-run 12 before any PRODUCTION_FLAG flip.

Producer ten-minute path

Force aloud: not-list (2), license stickies (3), KEEP (4), Hello exists (7). Those four stop the most expensive fork myths. Still finish the rest before production language.

Package table reminder (inside check 8)

CI sketch

# shipping-title — must stay green (check 4)
# brutal-lab — restore + Hello (checks 6–7)
# forbid: PRODUCTION_FLAG without check 12 sheet

Same-month noise (do not merge kits)

Opinion spine (quotable)

Stop forking BRUTAL over a live game because the mirror of Waylith looked finished. The free checks were always .NET 10, the not-list, license 2.1–2.3, Hello restore, and a signed KEEP — numbered so Discord cannot skip them.

Key takeaways

1. Twelve free checks before forking BRUTAL over a live title — primary keyword brutal framework checks.
2. Minimum gate: checks 1–4.
3. .NET 10 SDK literacy first.
4. README not-list: no editor / scene graph / renderer.
5. License 2.1–2.3 stickies before pride.
6. UNITY_KEEP (or equivalent) on the shipping game.
7. OS/GPU + NuGet restore on clean CI.
8. Hello window before any “we migrated” claim.
9. Package intent, not kitchen sink.
10. Owned spike beyond Hello.
11. Third-party LICENSES still apply.
12. Comms redlines + fork-over-live gate — PRODUCTION_FLAG off until signed.

FAQ

What are free BRUTAL framework checks?

Time-only gates on the public README, LICENSE, NuGet Hello path, and shipping-engine KEEP — plus this numbered kit. Sources: README, Ahwoo BRUTAL.

What is the four-check minimum?

.NET 10, not-list, license 2.1–2.3, shipping KEEP — checks 1–4.

Is forking free?

Cloning is free. Your time, middleware licenses, and credit obligations are not “free of work.”

Why not replace Unity on day one?

Because BRUTAL is a framework without Unity’s editor/content pipeline. See Stop Treating BRUTAL as Your Unity Replacement.

Where is the capability map?

What the BRUTAL C# Framework Changed for Indies - 2026.

Do I need Vulkan for check 7?

No. Hello window uses Glfw without a graphics API. Vulkan matters when you render (checks 5 and 9).

What about Mac?

Product page lists Mac OS among package targets; README getting-started emphasizes Windows/Linux. Verify current docs before promising parity (check 5).

Is this the license lab?

No. This is the numbered fork kit. Deep 2.2 orientation starts at the BRUTAL guide.

Can I skip Hello if I already know Vulkan?

No. Check 7 still proves NuGet + Glfw bootstrap on this machine/CI. Skip it and you learn about restore failures after you already promised production.

How does this relate to Unity fee fear?

Fee anxiety belongs in the engine-risk spreadsheet. This kit does not replace that sheet — it prevents a panic fork from pretending to be a plan.

Lab doc skeleton

Date:
Owner:
Check1 .NET10:
Check2 not-list:
Check3 2.1-2.3:
Check4 KEEP:
Check5 OS/GPU:
Check6 restore:
Check7 Hello:
Check8 intent:
Check9 spike:
Check10 third-party:
Check11 comms:
Check12 gate:
PRODUCTION_FLAG?: Y/N

Time-box table

If your estimate is “fork tonight and ship Steam tomorrow,” you are planning theater.

Projection: what “fork” means in this kit

People say fork when they mean three different things. Separate them before check 12:

Ahwoo’s public repo is the source of packages and discussion. A silent hard-fork that never contributes upstream is a company decision — name it, do not hide it inside “we starred BRUTAL.”

CONTRIBUTING / scope check (pairs with check 2)

Before you open issues demanding an editor:

1. Skim CONTRIBUTING.md scope notes in the repo.
2. Expect rejection of scene-graph / general-purpose engine drift — README already says so.
3. Put “we will build X in-house” on the check 8/9 intent list instead.

AI-assisted PR rules and sign-off expectations also live in contributing docs — read them before flooding the queue.

Security and conduct bookmarks

README points at SECURITY.md (private reports) and CODE OF CONDUCT. For a studio diligence folder, store those URLs next to check 3. Not glamorous — still part of “we adopted this stack” honesty.

NuGet profile sanity (check 6 deep)

Packages publish under Ahwoo’s NuGet profile (Ahwoo.Brutal.* in README samples). On restore failure:

1. Confirm package ID spelling against current README.
2. Confirm feed access (nuget.org vs private mirror).
3. Confirm TLS/proxy interception is not rewriting HTTPS.
4. Only then blame “BRUTAL is broken.”

Render spike checklist (inside check 9)

Minimum honest spike:

1. Window from check 7 still opens.
2. One clear color or triangle via a Vulkan (or chosen) path you can explain.
3. One failure mode documented (validation layer error, missing ICD, etc.).
4. Time spent recorded — if the spike ate a week with no triangle, KEEP shipping engine without shame.

Hiring note for check 12

Do not post “Unity refugees only, BRUTAL native” until checks 4 and 12 are real. Prefer “C# interop + Vulkan; Unity experience welcome; BRUTAL is evaluation.” Contract SOWs should mention KEEP and logo splash so compliance cannot be “out of scope.”

Support macros (check 11)

We keep the shipping engine. BRUTAL is a side prototype.
BRUTAL has no editor/renderer — we build what we need.
If we ship a BRUTAL-linked build: credits + logo splash + NOTICE.txt.

Group / org addendum (check 4 deep)

Second-machine restore smoke

After check 6 passes on CI, clone Hello on a second laptop. Record pass/fail. Cached-global-NuGet lies are how “it worked Friday” becomes Monday outage.

What this listicle is not

It is not a rewrite of what-changed. It is not the Unity Stop. It is not legal advice. Use those for narrative and refusal. Use this URL when someone asks for a numbered free check kit before a fork.

Where to leave the live flag

Keep PRODUCTION_FLAG_BRUTAL off until the twelve cards are honest. Point curiosity at the what-changed map and the Unity Stop. Orient compliance on the BRUTAL guide. Paste the Discord brief if your team only reads one message. Re-run check 3 the morning you cut a public BRUTAL-linked build.