# .gitignore for a make-my-arg game. It ships inside assets/starter/, so it lands in the project
# root when the starter tree is copied at scaffold time (step 6) — no rename, no second copy step.
# It keeps development by-products out of the repository while everything the deploy needs
# (index.html, pages/, assets/, data/*.json, tools/, viewer/) stays tracked.
#
# Two different jobs, two different files: .gitignore keeps junk out of Git history, .dockerignore /
# the rsync exclude list keeps development material out of the deployed image. A plaintext table must be
# listed in both (it is worse than junk: it is the answer key, R3).

# ── Development-only plaintext sources (R3): the answer key, never committed, never deployed ──
data/*.src.json

# ── dev/ops trees: no site page lives here (the tools skip them via CONFIG.skipDirs) ──
docs/site-graph*
memory/
review_comments.md
*.log

# ── Story source material (may be copyright-protected; keep it locally, do not publish it) ──
docs/source-material/

# ── OS files ──
.DS_Store
Thumbs.db

# ── Editor / IDE ──
.vscode/
.idea/

# ── Agent scratch dirs (an agent working in this repo writes its own logs here) ──
.codex/
.opencode/
.claude/
.superpowers/
docs/superpowers/

# ── Node (the tools need no install; a project that adds one gets node_modules) ──
node_modules/
