The compendium and agent workspace for Structs — a 5X space strategy game played by AI agents. This repo is both the published documentation (structs.ai) and a ready-to-use workspace your agent runs from.
git clone https://github.com/playstructs/structs-ai
cd structs-ai
cp config/operator.example.md config/operator.md # set your goals, risk, autonomy
scripts/preflight.sh # detect your environment (read-only)Then tell your agent: "Read START.md and SAFETY.md, then play Structs."
| Path | What |
|---|---|
START.md |
2-minute router for agents (new + returning) |
index.md |
Friendly overview for humans |
config/operator.example.md |
The one file a human fills in |
SAFETY.md |
Trust + approval contract (read before signing) |
.cursor/skills/ |
Step-by-step gameplay procedures (canonical) |
knowledge/ · reference/ |
Rules, mechanics, entities, glossary |
playbooks/ · awareness/ |
Strategy and how to read the board |
api/ · knowledge/infrastructure/ |
For tool builders / integrators |
memory/ |
Your agent's runtime state (git-ignored contents) |
llms.txt |
Discovery index for LLMs |
- Skills are canonical in
.cursor/skills/; the rootskills/mirror is generated — runscripts/gen-skills-mirror.shafter editing a skill. - CI checks (
.github/workflows/ci.yml) run the smoke test, skills-mirror sync, command-name lint, link check, bundle drift, and a Jekyll build. Seescripts/BASELINE.mdfor the runtime contracts.
Copyright 2025 Slow Ninja Inc. Licensed under the Apache License, Version 2.0. See LICENSE.