Learn the foundations.
Choose the format that works for you, then build a small original project.
THE CREATOR TOOLKIT
Understand the system.
Build something yours.
A workbook, a repeatable research method, and a small example you can run. Start with evidence. Keep your implementation original.
01 / A SMALL, WORKING START
Your first
validated matrix.
Run the original example before you model a real project. The starter converts the Entities, Fields and Rules sheets into JSON. It does not reconstruct games or prove behavioral parity.
Before you start
- Git and Python 3.10 or newer installed.
- A terminal and a new working folder. If
rustportsalready exists, use a different folder rather than overwriting it. - Excel or LibreOffice only if you want to edit the workbook.
- No game, server, GitHub login or Rust compiler needed for this Python example.
Trusted local workbooks only. This experimental validator is not a sandbox or a public upload service. Review code before running it and never include secrets or third-party game files.
Read the starter documentation ↗Windows · Command Prompt
git clone https://github.com/phoenixfire808/rustports.git
cd rustports\toolkit\spreadsheet
py -3 -m venv .venv
.venv\Scripts\python -m pip install -r requirements.txt
.venv\Scripts\python scripts/build_matrix.py
.venv\Scripts\python scripts/build_matrix.py --check
.venv\Scripts\python -m unittest discover -s tests -vmacOS / Linux · Terminal
git clone https://github.com/phoenixfire808/rustports.git
cd rustports/toolkit/spreadsheet
python3 -m venv .venv
.venv/bin/python -m pip install -r requirements.txt
.venv/bin/python scripts/build_matrix.py
.venv/bin/python scripts/build_matrix.py --check
.venv/bin/python -m unittest discover -s tests -vWhat success looks like
You get generated/matrix.json with five original example rows. The --check command exits successfully without rewriting it, and the test suite reports OK.
Then copy the starter to your own project, preserve the example, and edit your copy of workbooks/model.xlsx. Fix data in the workbook, not in generated JSON.
If py is unavailable on Windows, use your installed Python 3.10+ executable. On Linux, your distribution may require its Python venv package. Do not use administrator privileges to work around dependency errors.
02 / SPREADSHEET-FIRST RESEARCH
Make the unknowns visible.
Reverse engineering starts with careful observation, not confident guesses. Work only with systems you are authorized to analyze. Keep evidence, hypotheses and implementation separate.
- 01 / OBSERVE
Record the evidence.
Define scope and permissions. Log the setup, action, observed result, source and confidence. Repeat the observation.
- 02 / MODEL
Give facts stable IDs.
Describe entities, fields and rules. Keep unsupported hypotheses in Evidence, not in runtime constants.
- 03 / GENERATE
Validate before building.
Export the matrix. Resolve ID, type and reference errors in the sheet. Let generic engine code consume it.
- 04 / VERIFY
Test the real behavior.
A valid matrix is not a correct game. Compare observable runtime behavior with your recorded reference and document differences.
03 / CHOOSE YOUR LEARNING PATH
Less link collecting.
More forward motion.
Use the right reference.
Find official documentation for Ghidra, radare2, debuggers and network analysis. Use tools only within your authorization.
Open the reference library ↗No proprietary binaries, extracted assets or circumvention recipes are distributed here.
Know the schema.
Look up supported sheets, columns, IDs, field types and references. See what the validator checks and what it does not.
Read the matrix reference ↗Creator-run connection guide ↗SQLite, DBViewer & Obsidian.
ZIP includes JcodeSS, separate DBViewer, launchers, licenses and checksums. Unoptimized preview, not a stable release. Open one Obsidian vault for automatic memory-boundary sync.
Learn the single-source-of-truth workflow, workbook cutover boundary, local memory sync, token-efficiency limits and end-to-end validation flow.
Open the detailed guide and flowchart ↗04 / TRANSPARENT BY DEFAULT
Useful today.
Still being built.
We hope to complete the broader workflow soon. There is no committed release date, and the native Jcode draft is not a stable release.
Python workbook starter
Editable synthetic workbook, deterministic JSON export, validation, non-writing check mode and tests. Trusted local input only. It is not a game engine or a hostile-file sandbox.
Native Jcode integration
The source snapshot is public for review. CLI dispatch, reference parsing and resource limits need work. No tested binary or stable native release is supplied with that legacy snapshot. A separate Windows JcodeSS preview now has five passing Obsidian sync tests and packaged CLI/SQLite smoke checks, but full interactive acceptance remains unverified. Its schema differs from the Python starter. Read the blockers before trying it.
Verify the complete workflow
Native cross-platform tests, stronger evidence validation, and an original Rust runtime example that consumes the matrix. See the researched 64-item priority roadmap.
05 / BUILD WITH US
Small contributions.
Real leverage.
Reproduce a bug with the synthetic workbook. Improve a confusing step. Add a focused regression test. Share your environment and evidence, not secrets or third-party game files.
Report a bug or propose an improvement ↗Keep rights and responsibilities clear. Tool availability is not permission to analyze or redistribute someone else’s work. RustPorts does not promise zero legal risk. Read the terms and contact us about rights concerns.