Hand-written like items/ — no build-tool namespace needed. Blurbs are content;
mechanics are code, and a test asserts the blurb files carry no hit dice. The
prose is placeholder; swapping it later touches no code, which is what stops the
content BOM from blocking the engine.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
Mechanics are rules and live in code, beside Luck and Currency. The skill pools
are new — the races/classes spec said only 'the stat-appropriate slice' and never
enumerated them; each pool is larger than its pick count so the choice is real.
The dwarf's +2 vs poison is exposed separately from save() rather than folded
into it: baking a conditional into a flat number is how a sheet starts lying.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
The tag extractor returns every occurrence of a move tag and the validator is
a pure membership test, so MoveApplier applied duplicates. A model repeating a
tag N times therefore picked an amount in unary — the AI choosing a number,
which is the §2 breach the bounded vocabulary exists to prevent.
The currency fix (8e1238c) closed this for denominations only. Generalizing it
to every move that names a thing covers real items too — three give_item(amulet)
tags are one amulet, not three — and lets the currency special case be deleted
rather than kept alongside a second mechanism. ID_MOVES is reused from
MoveValidator instead of restated.
Also caps adjust_disposition on the reply's NET swing. MAX_DELTA was clamped
per tag, so three +15s moved standing by 45 — precisely the wholesale swing its
own comment says it forbids. Applied in emission order, so a later
become_hostile still floors what came before it.
offer_quest and reveal needed no change: add_quest self-guards on has_quest and
mark_revealed is idempotent.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
The existing test varied the move name and the denomination together, so a
name-only or denom-only key would have passed it too. Two assertions nail it:
give_item(gold)+give_item(silver) kills a name-only key, and
give_item(gold)+accept_item(gold) kills a denom-only key.
The spec now records the per-reply ceiling as a decision rather than an
accident, and flags the pre-existing non-currency duplicate-tag path as open.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
TagExtractor collects every [MOVE:] tag occurrence with no dedupe, and
MoveValidator is a pure membership test against available_moves, so N
copies of one legal move all pass validation. MoveApplier then looped
and applied every copy. A model could therefore pick an arbitrary
amount of money in unary by repeating one tag: [MOVE: accept_item(copper)]
x47 against a 47c purse drains it to 0; [MOVE: give_item(gold)] x10
mints +100,000c. The eight-move vocabulary (§6) takes no quantity
argument precisely so the AI cannot choose a number — this closed that
gap by another door and was a §2 breach (AI owns text, never state).
MoveApplier.apply() now tracks currency moves (give_item/accept_item on
a denomination) already applied within one call, keyed on name+denom,
and skips repeats. Non-currency moves are untouched — repeated
give_item(amulet) still yields multiple copies, which stays bounded by
the pre-existing cross-reply gifts_given gate and is a separate,
already-reported issue. move_validator.gd and tag_extractor.gd are
untouched.
Also adds the Finding-2 test the currency spec (§7) named but never
delivered: after GameState.grant("gold", 1), inventory stays empty, so
accept_item(gold) in available_moves comes only from the affordability
loop, never double-offered via the inventory.keys() loop.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
The seed placeholder was a flat 1240 gold — 12.4 million copper under the real
model, which is 1,240 plot points in the command bar. It reseeds to 347c, a
party carrying silver and no gold: '◈ 3s 47c'.
This is the only call site of Currency.format() — the display edge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
The three denominations are move tokens, not stored buckets — there is one
integer, so change-making never arises. The deserter is down to 47c: under a
silver, so he cannot cover a bed and the player sees it on turn one.
A parity test guards the code-side ratios against the content-side ids.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
give_item(gold) is +10,000c and does NOT mark gifts_given — that gate is a
global one-shot keyed by item id, so marking money would let an NPC pay you
exactly once per campaign. accept_item(<denom>) is offered only for the
denominations the purse can actually pay.
The eight moves do not grow. MoveValidator is untouched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
One int holds all money. grant/take are the sole routing point — currency ids
go to the purse, everything else to inventory. Money is never an inventory row.
The 'coin' stand-in in the two unrelated inventory tests becomes
worn_shortsword; they were never about money.
One integer, stored in copper. The ratios are rules and live in code; prices
are content and land in M8. format() suppresses empty denominations so being
broke reads as '47c', not '0g 0s 47c'.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QYa9u7Kdxv5gX4AnwWexy8
The purge plan assumed content/world/** was pure build output. The client
reads it: content_db.gd loads content/world/topics/ (now absent, since every
topic was Duncarrow-generated) and its tests asserted on deleted ids.
- content_db: a missing content dir is an empty dir, not an error (§13)
- test_content_db: repoint at cosmology canon; drop deleted npc.mera-fenn
- test_content_db: assert the no-body-on-client invariant structurally, so it
cannot pass vacuously when the topic set is empty
- schema.md: finish repointing the worked example onto specimen.md's ids
- restore a local real-bible --check smoke test (was only in CI)
- roadmap: record that generated content has client consumers
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The scripts were tracked but their .gd.uid siblings weren't. Godot 4.4+ .uid
files pin each script's uid://, so leaving them out lets each clone regenerate
different uids and break uid:// references in scenes/resources. Matches the 159
.uid files already tracked in the repo.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HuHRPE7VfppUJEaoGBEUqZ
The hex-removal correction set _glow.color = Palette.BLOOD (opaque), which
with the 0.5-0.8 modulate pulse read as a near-solid red slab. Restore the
intended low alpha via Color(Palette.BLOOD, 0.22) — RGB still from Palette,
alpha a numeric param (same idiom as the ember ramp). Whole-branch review
catch. Title tests 5/5.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Add scripts/sync-version to sync root /VERSION into client/project.godot
- Add Version GDScript helper (reads application/config/version)
- Test coverage: 2 tests for Version.string() and footer()
- Synced config/version="0.01-alpha" to project.godot
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Capture the direction change from M3-b: UI screens are authored editor-first
as .tscn node trees (previewable/arrangeable in the Godot editor), scripts do
on-load work only — never build the tree in _ready(). Add ADR 0001, a
"UI scene conventions" how-to in client/docs, a §16 mandate in the charter,
and mark M3-a/M3-b done in the roadmap.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wider consumable tiles overflowed the command bar, shoving END TURN under
the dock column. Trim the vitals min-width (230->120) and END TURN
(150->140) and separation so the row fits — END TURN lands cleanly at the
bottom-right, below the dock.
Move the dock handle to the outer (right) side per mock 2a: the menu now
sits left of the handle and slides right off-screen to hide, the handle
staying at the world's right edge. Dock sits flush to that edge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The 62px tiles clipped longer names (Draught, Antidote). Widen to 93px
(~1.5x) — there is room in the command bar — so labels fit while tiles stay
uniform.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The narration RichTextLabel had no theme styling, so it rendered near-white
and was unreadable on parchment. Add a base RichTextLabel style to the theme
(dark INK_BODY serif at 20px, italic face wired) — systemic and previews in
the editor; every parchment prose surface now reads as dark ink. Also clip
the consumable slot labels so a long name ("Antidote") can't stretch its tile
wider than the empty ones — all 8 stay a uniform 62px.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move the three shell scenes (SystemDock, NarrationBook, MainWindowShell)
from build-everything-in-_ready() to authored .tscn node trees with the
theme + type-variations set in the editor, so the layout previews and can
be arranged in the Godot editor. Scripts keep only on-load work: @onready
node refs, DM-loop wiring, the dock slide, and binding seed ShellState into
the fixed authored turn-order tokens + consumable slots.
Fixes the dock offset + over-long buttons (now fixed-width DockButtons in a
right-anchored menu). Headless probe: root 1920x1080, world/book split
1180/740, dock at the world's right edge — no collapse. Suite 138/138.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ThemeKeys.PARCHMENT_CARD is a PanelContainer-base variation (only a
`panel` stylebox) — a Button looks up normal/hover/pressed/disabled,
which don't exist there, so a Button styled with it renders unskinned.
Add PARCHMENT_BUTTON, a Button-base variation with those four
styleboxes plus an empty focus box, built from Palette colours only.
Same fix pattern as DOCK_BUTTON for the system dock. Regenerated
game_theme.tres from the builder.
DarkPanel is a PanelContainer-base variation (only a `panel`
stylebox); a Button looks up normal/hover/pressed, which don't exist
on it, so SystemDock's menu buttons rendered as default gray Godot
buttons instead of mock-faithful dark nav buttons.
Add a dedicated Button-base DockButton variation (Palette-only
colours, dock-dark normal/hover/pressed styleboxes), point
system_dock.gd's menu buttons at it, and regenerate the committed
game_theme.tres artifact from the builder.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Keep the design/logical canvas at 1920x1080 (charter §16 / mockups) but add a
window-size override so the window opens at 1600x900 and doesn't fill the whole
screen; canvas_items + keep-aspect scales the canvas to fit.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two root causes behind the blank F6 run:
1. project.godot had no display/stretch config (README mandates canvas_items
@ 1920x1080). Added [display] window size + stretch mode/aspect — the
project's canvas foundation, needed by every screen.
2. A root Control run directly (F6) is not sized to the viewport by full-rect
anchors — it collapses to (0,0), clipping the ScrollContainer and hiding all
content. Size the root to the viewport explicitly and track size_changed
(dropping the full-rect anchors, which conflict with setting size directly).
Faithful main-scene run now measures root = scroll = viewport = 1920x1080.
Suite 125/125.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
palette.gd / theme_keys.gd already declare `class_name`, so `const Palette =
preload(...)` in every consumer shadowed the global class and emitted 4
editor errors ("has the same name as a global class"). Reference the global
class_names directly. Editor reload now clean; suite 125/125; runtime probe
still builds all 75 showcase nodes with the theme applied.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- build_game_theme.gd: extract build_theme() as a static func returning the
fully-built Theme; _init() is now a thin save+quit wrapper. Enables a test
to build a fresh Theme in memory without running the SceneTree main loop.
- test_theme_resource.gd: add test_committed_tres_matches_builder() — builds
a fresh Theme via the preloaded builder and compares bg_color against the
committed game_theme.tres for PrimaryCTA/normal, ParchmentCard/panel, and
DarkPanel/panel. Catches "palette/builder changed, .tres not regenerated."
- build_game_theme.gd: give Tab, TabActive, and Chip explicit hover/pressed/
focus styleboxes (focus is StyleBoxEmpty to suppress the default ring) so
they no longer fall back to Godot's default gray button on interaction.
Add a disabled stylebox + font_disabled_color to PrimaryCTA for the
unaffordable-CTA state the 2a shell will need.
- theme_keys.gd: add HEADING/ACCENT/MONO font-role consts (kept out of ALL,
which is stylebox-variation-only); build_game_theme.gd and
theme_showcase.gd now reference them instead of raw StringName literals.
- theme_showcase.gd: add an italic DM-voice RichTextLabel (EB Garamond
italic face, spec §6) and two semantic-coloured chips (BLOOD for
aggressive actions, GOLD for pay/faith) alongside the existing neutral
chip, demonstrating the script-applied colouring from spec §5.3.
- dark_bay.gdshader: comment noting the stripe uses FRAGCOORD (screen-space)
while the vignette uses UV (node-local) — intended for full-bleed use.
- surfaces/*.gd: ## comments documenting the accepted deviation from spec
§5.4 (uniforms set at runtime from Palette rather than baked into the
.tscn) — no behavior change.
- Regenerated game_theme.tres via build_game_theme.gd.
Full suite: 125/125 (was 124, +1 drift-guard test).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GFDm1ku9WDUacK2S78m2be
- _surfaces(): add VignetteOverlay.tscn alongside ParchmentPanel/DarkBay,
backed by a Palette.STEEL ColorRect so the dim is visibly rendering.
- _variations(): add ItemTileEmpty (next to filled ItemTile), ParchmentInset,
and DarkPanel stylebox nodes.
- _fonts(): new section rendering default serif body, Accent
(Architects Daughter), and Mono (JetBrains Mono) labels at real sizes.
- _swatches(): wrap each ColorRect in a VBoxContainer with a Mono caption
Label naming its Palette const, per spec's "labelled swatches."
No raw hex / Color(...) literals introduced; all colour via Palette.*, all
styleboxes via ThemeKeys.* variation names.
By-hand cold-start test showed 3 phrases before the response; 1.5x the
cadence so each line stays on screen a touch longer.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The throwaway harness hand-built a minimal canon log and left luck_descriptor
as the default "", which fails the canon-log schema (minLength 1) → /npc/speak
422s and the harness silently degrades to fallback. Build the player through
LogPlayer with a non-empty §7 fortune line. Verified: request now 200s and Fenn
answers in-voice.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
