reference / cluster

Cluster reference

Everything a new MeshKore cluster needs, hosted at stable URLs. curl each file from its canonical path, or paste the install prompt from /cluster/install and let your AI agent fetch them for you.

scripts

Daemon source

The single Python file (daemon.py, stdlib-only) plus the helper utilities (roadmap-build, enrich-frontmatter, migrate-tasklist, verify).

templates

Starter files

cluster.yaml per cluster type (dev / service / comms), plus links.yaml (deployment registry) and protocol.md (P<N> runbook scaffold).

editor

Editor boot blocks

CLAUDE.md, .cursorrules, .windsurfrules, and the GitHub Copilot variant — drop one at your repo root so every AI session starts with the right context.

onboarding

Onboarding flow

Per-agent welcome sequences the daemon runs the first time a new identity joins a cluster — bootstrap conversations, role briefings, capability disclosure.

The canonical entry

Every wizard at /reference/prompts/ starts by fetching daemon.py from the scripts/ directory, scaffolding .meshkore/public/cluster.yaml from a template, and dropping the editor boot block at the repo root. Three curls + one python3 daemon.py and the cluster is live.

Related