Skip to content

Latest commit

 

History

History
73 lines (56 loc) · 3.64 KB

File metadata and controls

73 lines (56 loc) · 3.64 KB

SiliconSaga Realm — AI Agent Context

This realm declares the SiliconSaga ecosystem components, identity, and component-specific skills. It is loaded automatically when present in realms/realm-siliconsaga/.

For workspace-level instructions (GDD methodology, ws CLI, git workflow, auth setup), see the root AGENTS.md.


Repo Roles (Community Components)

Repo Tier Role Path
nordri 1 Cluster substrate (Traefik, Crossplane, Velero, ArgoCD) components/nordri
nidavellir 2 Platform app-of-apps (Vegvísir, Mimir, Keycloak, …) components/nidavellir
mimir 2 component Data services via Crossplane + operators components/mimir
vordu 2 component BDD roadmap visualization components/vordu
heimdall 2 component Observability stack components/heimdall
tafl 2 Board game engine service components/tafl
bifrost 2 Bridge/gateway service components/bifrost
ymir 3 End-user platform components/ymir
terasology 3 Voxel game (fork) components/terasology
destinationsol 3 Space shooter game (fork) components/destinationsol

Realm Skills

Skills in this realm live in .agent/skills/<name>/SKILL.md (relative to this realm directory). They are discovered automatically by the GDD orientation skill during trust verification.

Per the SiliconSaga skill taxonomy design (docs/plans/2026-05-30-skill-taxonomy-design.md, revised after first-pass review), opinionated platform-choice skills (ArgoCD, Crossplane, AlertManager, kube-prometheus-stack) live in their owning components: components/heimdall/.agent/skills/ and components/nordri/.agent/skills/. The realm carries the navigation index (siliconsaga-stack) and realm-internal skills (e.g. terasology-testing).

Skill Name Description Target Component(s)
siliconsaga-stack Stack narrative + skill index — which component owns which capability, the alert pipeline end-to-end, GitOps model (argo ns + in-cluster seed-Gitea), homelab-vs-GKE seam, bare-workspace recipe realm-wide
terasology-testing Engine-level and MTE integration test patterns, network event gotchas, Gradle execution terasology

Tone Guide

While most repos can follow any communication style that seems appropriate, some are user-facing in sensitive areas:

  • demicracy - next generation civics platform, likely tech-skeptical audience focused on local community topics
  • schools - parents, union members, school board, etc - may contain highly sensitive content that risks conflicts between groups
  • mtl-site - sports volunteers, not likely very technical, needs minimal jargon and low barrier to entry

When writing anything that might reach these groups:

  • Avoid techno-utopian vocabulary. "Post-scarcity," "cyber," "decentralized," "crypto," "DAO" all read as cult-of-Silicon-Valley to the audience and poison the message before it lands.
  • Prefer grounded, service-oriented language. "Gentle cost reduction," "community self-sufficiency," "overhead removal," "coordination tools."
  • Avoid "Demicracy" branding until live. Talk about "coordination tools" or "community platform" in the abstract. The brand name is fine on demicracy.org, which is aimed at people who opted in.
  • The audience is not platform engineers. Parents and board members are the primary readers. Board members skim. Parents are tired.

This distinction between public-facing civics content and internal platform/infra work matters — other components can use the normal engineering vocabulary because the audience opted in.