GroundMesh builder start

Start small, inspect first, then add one careful change.

GroundMesh is the canonical core repo for the Atlas-backed docs layer, contributor flow, governance artifacts, and public trust scaffolding. The newer groundmesh-world site is the stronger narrative seed, and this repo is where we keep the operational memory clear and extend it carefully.

Open Atlas Contribute Front Door Landscape
What you need

Basic local setup

  • Windows PowerShell
  • Git
  • Python available on PATH for JSON and helper scripts
  • A willingness to work in small batches
Reality check

GroundMesh and GroundMesh World

Treat GroundMesh as the canonical core and groundmesh-world as the public narrative seed. We join them by translation and careful reuse, not by forcing a hard repo collapse before the structure is ready.

One home, one spine

Use one active cockpit and one shared memory

GroundMesh works best when active repo work happens in one main working surface and important state is anchored in GitHub rather than scattered across many chat windows.

  • Use Codex for edits, checks, and branch-sized implementation batches
  • Use GitHub as the source of truth for files, PRs, issues, and decisions
  • Use ChatGPT for reflection, review, and strategy when broader thinking is needed
Reduce drift

Do not repeat the same state everywhere

If something matters, anchor it once in GitHub and let the other surfaces point back to that record instead of recreating the same context again and again.

  • Move real decisions into repo docs, ADRs, Atlas, PRs, or issues
  • Use chat for translation and handoff, not as the only memory
  • Prefer one clear next step over many parallel promises
ChatGPT bridge

Import old conversations without living on copy-paste

If you export your ChatGPT data, GroundMesh can ingest the resulting conversations.json or export zip into a local archive for later reading, indexing, and careful reuse. 

.\scripts\import-chatgpt-export.ps1 -SourcePath C:\path\to\chatgpt-export.zip
What it creates

Readable local history

The importer writes a timestamped folder under archives/chatgpt_exports with a raw copy of the export, a top-level index, a manifest, and one Markdown file per conversation.

1

Clone and bootstrap

git clone https://github.com/mailgmirko-creator/groundmesh.git
cd groundmesh
Set-ExecutionPolicy -Scope Process Bypass -Force
.\scripts\bootstrap-node.ps1
2

Check what already works

.\scripts\health-check.ps1
.\scripts\atlas-generate.ps1
3

Open the current map

start .\docs\atlas\index.html
start .\docs\index.html
First PR loop

Make one small change and ship it cleanly

git checkout -b docs/my-small-change
# edit one page, doc, or script
.\scripts\atlas-generate.ps1
.\scripts\health-check.ps1
git status
git add docs\path-you-changed
git commit -m "docs: explain the small change"
git push -u origin docs/my-small-change

Then open GitHub, create the PR into main, and keep the title and body honest about what changed and why.

Good first batches

Start with one visible improvement

  • Fix one broken link or confusing sentence
  • Clarify one public page without changing the whole tone
  • Add one Atlas entry for something that already exists
  • Tighten one workflow step or one trust note
  • Avoid mixing design, policy, and structural edits in one batch
First safe contribution path

Use Atlas before adding structure

  • Read the Atlas registry and current generated page
  • Inspect the page or script you want to touch
  • Make one small batch of edits
  • Update status or registry if the public state changed
  • Regenerate Atlas and rerun health checks