Merge canon log schema design + implementation plan

Design spec + Plan A (contract & api validation) + content/world reorg.

.claude/security/log.txt

@@ -1,46 +0,0 @@
-[2026-07-09 10:54:03.926] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:54:03.929] Processing: hook_event=UserPromptSubmit, tool=
-[2026-07-09 10:56:09.126] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:09.128] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:18.474] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:18.476] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:21.852] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:21.855] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:25.336] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:25.338] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:28.741] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:28.744] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:32.070] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:32.073] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:35.419] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:35.422] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:39.157] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:39.159] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:42.516] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:42.518] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:48.542] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:48.545] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:51.814] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:51.816] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:55.095] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:55.098] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:56:59.075] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:56:59.077] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:02.336] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:02.338] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:05.616] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:05.618] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:08.886] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:08.888] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:12.207] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:12.209] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:16.216] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:16.218] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:19.540] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:19.542] Processing: hook_event=PostToolUse, tool=Write
-[2026-07-09 10:57:31.383] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:31.386] Processing: hook_event=PostToolUse, tool=Edit
-[2026-07-09 10:57:51.755] Hook called with args: ['/home/ptarrant/.claude/plugins/cache/claude-plugins-official/security-guidance/2.0.6/hooks/security_reminder_hook.py']
-[2026-07-09 10:57:51.758] Processing: hook_event=Stop, tool=
-[2026-07-09 10:57:51.761] _git_status_porcelain error: [Errno 138] emscripten does not support processes.
-[2026-07-09 10:57:51.761] Stop hook: empty review set

.claude/security/security_warnings_state_1a541ba7-dd07-4b76-b6ff-9f6ad45bfc8e.json

@@ -1 +0,0 @@
-{"shown_warnings": [], "touched_paths": []}

.gitignore

@@ -0,0 +1,63 @@
+# ─────────────────────────────────────────────
+# Godot 4.7 (client/)
+# ─────────────────────────────────────────────
+# Editor + import cache (4.x replaced .import/ with .godot/)
+.godot/
+# Exported builds
+/client/build/
+export.cfg
+# export_presets.cfg holds keystore paths / signing config — keep secrets out of history
+export_presets.cfg
+# Mono/C# (only if we drop to C# per charter §16)
+.mono/
+data_*/
+*.mono/
+
+# ─────────────────────────────────────────────
+# Python / FastAPI (api/)
+# ─────────────────────────────────────────────
+__pycache__/
+*.py[cod]
+*$py.class
+*.egg-info/
+.eggs/
+build/
+dist/
+# Virtual envs
+.venv/
+venv/
+env/
+ENV/
+# Tooling caches
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+htmlcov/
+.tox/
+
+# ─────────────────────────────────────────────
+# Secrets — the API key lives here in dev, NEVER in the client (charter §4)
+# ─────────────────────────────────────────────
+.env
+.env.*
+!.env.example
+
+# ─────────────────────────────────────────────
+# Editors / OS
+# ─────────────────────────────────────────────
+.vscode/
+.idea/
+*.swp
+.DS_Store
+Thumbs.db
+
+# ─────────────────────────────────────────────
+# Logs (charter §4/§10 logs go to a store, not the repo)
+# ─────────────────────────────────────────────
+*.log
+
+# ─────────────────────────────────────────────
+# Claude Code harness session/state files
+# ─────────────────────────────────────────────
+.claude/

CLAUDE.md

@@ -530,3 +530,15 @@ Test that before building anything else. Every other system exists to make that
 - **When a request conflicts with this document, say so.** Do not silently comply. This file is the argument; if it is wrong, change the file first.
 - **When adding a system, state which side of §2 it falls on.** State or text. If it is both, it is two systems.
 - **Prefer deleting scope.** The POC's value is answering one question fast.
+
+### Git workflow
+
+The standard flow for this project. Claude follows it; the human owns `master`.
+
+- **Branches.** `master` and `dev` are long-lived. `master` is release; `dev` is integration. All work branches off `dev`.
+- **Never commit non-doc changes directly to `master` or `dev`.** Anything that is not purely docs gets its own branch (`feature/…`, `fix/…`, `chore/…`, etc.), branched off `dev`.
+- **Doc-only edits may commit directly to `dev`** — no branch required.
+- **Never commit to `master`.** Ever. The human merges `dev` → `master` as they see fit. Claude never touches that merge.
+- **Merging a working branch → `dev`** happens only after the work is done, smoke-tested, and **the human confirms**. Merge locally with `--no-ff`.
+- **After merging to `dev`, delete the working branch locally.**
+- **Pushing.** Origin is set. Claude pushes **only `dev`**, and only after a merge. Never push working branches or `master` — the human always pushes `master`.

api/.dockerignore

@@ -0,0 +1,16 @@
+# Keep the build context small and secrets out of the image.
+.env
+.env.*
+__pycache__/
+*.py[cod]
+.venv/
+venv/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+htmlcov/
+docs/
+README.md
+Dockerfile
+.dockerignore

api/.env.example

@@ -0,0 +1,11 @@
+# Copy to .env for local dev. NEVER commit the real .env. The client never sees
+# any of these — keys live only in the proxy (charter §4).
+
+# Where the proxy sends model calls in dev (Ollama on the homelab).
+OLLAMA_BASE_URL=http://localhost:11434
+
+# Prod model provider (charter §4). Leave blank in dev.
+REPLICATE_API_TOKEN=
+
+# Port the proxy binds. compose / fly.io override this.
+PORT=8000

api/Dockerfile

@@ -0,0 +1,26 @@
+# /api — FastAPI proxy (charter §4). Docker-based for parity across
+# dev / homelab VPS / fly.io.
+FROM python:3.12-slim
+
+# Faster, quieter Python in containers.
+ENV PYTHONUNBUFFERED=1 \
+    PYTHONDONTWRITEBYTECODE=1 \
+    PIP_NO_CACHE_DIR=1
+
+WORKDIR /app
+
+# Install deps first so the layer caches when only app code changes.
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# App code.
+COPY app ./app
+
+# Run as non-root.
+RUN useradd --create-home --uid 1000 appuser
+USER appuser
+
+EXPOSE 8000
+
+# fly.io / compose set PORT; default to 8000 (charter localhost:8000).
+CMD ["sh", "-c", "uvicorn app.main:app --host 0.0.0.0 --port ${PORT:-8000}"]

api/app/main.py

@@ -0,0 +1,47 @@
+"""FastAPI proxy entrypoint — the guarding proxy of charter §4.
+
+Skeleton only: a health check plus the five role endpoints as stubs so the
+container runs and the client has something to point at. Real prompt routing,
+model selection, logging, and auth land later — all server-side (§4, §5).
+"""
+
+from fastapi import FastAPI
+
+app = FastAPI(title="coc-rpg proxy", version="0.0.1")
+
+
+@app.get("/health")
+def health() -> dict:
+    """Liveness probe for compose / fly.io."""
+    return {"status": "ok"}
+
+
+# ── Role endpoints (charter §4) ──────────────────────────────────────────────
+# The client knows these paths and nothing about which model or prompt serves
+# them. Stubs for now; each will load its prompt from api/prompts/ and route to
+# the configured model.
+
+
+@app.post("/dm/narrate")
+def narrate() -> dict:
+    return {"detail": "not implemented"}
+
+
+@app.post("/dm/adjudicate")
+def adjudicate() -> dict:
+    return {"detail": "not implemented"}
+
+
+@app.post("/dm/improvise")
+def improvise() -> dict:
+    return {"detail": "not implemented"}
+
+
+@app.post("/npc/speak")
+def npc_speak() -> dict:
+    return {"detail": "not implemented"}
+
+
+@app.post("/party/banter")
+def banter() -> dict:
+    return {"detail": "not implemented"}

api/fly.toml

@@ -0,0 +1,21 @@
+# fly.io deploy config for the proxy (charter §4, prod). Stub — set app name and
+# region, then `fly deploy` from api/. Secrets (REPLICATE_API_TOKEN) go via
+# `fly secrets set`, never in this file.
+app = "coc-rpg-proxy"      # TODO: claim a real app name
+primary_region = "ord"     # TODO: pick a region
+
+[build]
+  dockerfile = "Dockerfile"
+
+[http_service]
+  internal_port = 8000
+  force_https = true
+  auto_stop_machines = true
+  auto_start_machines = true
+  min_machines_running = 0
+
+[[http_service.checks]]
+  method = "GET"
+  path = "/health"
+  interval = "15s"
+  timeout = "2s"

api/requirements.txt

@@ -0,0 +1,5 @@
+# FastAPI proxy runtime deps. Pinned from first resolve; bump deliberately.
+fastapi==0.139.0
+uvicorn[standard]==0.51.0
+httpx==0.28.1        # calling Ollama / Replicate
+pydantic==2.13.4     # request/response contracts

content/README.md

@@ -1,15 +1,37 @@
 # /content — authored game data
 
-Cross-cutting authored writing. Consumed by **both** sides: the client ships fallback text and quest/story data; the api reads NPC knowledge lists to build prompts. That shared ownership is why it sits at the repo root, not inside either folder.
+Cross-cutting authored writing, split by role in the data model (see the canon
+log spec in [`/docs`](../docs)):
 
 ```
-/quests    Story skeletons and quest definitions (authored, not AI-generated — §17)
+/world     Static, ID-referenced game content. Identical every playthrough.
-/npcs      Per-NPC knowledge lists — the entire content an NPC can draw on (§6)
+  /locations   The map — towns, dungeons, points of interest (by id)
-/fallback  Authored degraded-DM text for every AI surface (§13)
+  /npcs        Per-NPC persona + knowledge lists (charter §6)
+  /quests      Story skeletons and quest definitions (charter §17)
+  /items       Item definitions — gear, consumables, cursed items (§7)
+/origins   Thin starting-state seeds. One per starting point. POC authors one.
+/fallback  Authored degraded-DM text for every AI surface (charter §13)
 ```
 
+## The three layers
+
+- **`world/`** is static content the origin and the canon log reference by stable
+  string **id**. Authored once; the same for every play.
+- **`origins/`** are thin seeds — where the player starts, the situation, and
+  disposition/quest/item/build seeds. Values, not maps. Increase replayability
+  without multiplying authoring work.
+- **`fallback/`** is not world content — it is degraded-DM prose (§13), consumed
+  when the API is down. It sits outside `world/` on purpose: nothing resolves an
+  id against it.
+
+At new-game, code constructs the runtime **canon log** from a chosen origin +
+world content + character creation. See [`/docs/canon-log.md`](../docs) (spec) —
+authored via the design under [`/docs/superpowers/specs`](../docs/superpowers/specs).
+
 ## Authoring notes
 
-- **NPC knowledge lists are the whole design** (§6). They are the only thing stopping the blacksmith from revealing the twist. Real authoring work — budget for it.
+- **NPC knowledge lists are the whole design** (§6). The only thing stopping the
-- **Fallback text is content, not error handling** (§13). Written in the DM's voice, lives beside the rest of the writing. Every AI-dependent surface needs one before it ships.
+  blacksmith from revealing the twist. Real authoring work — budget for it.
-- Story skeletons are authored for now. AI-generated skeletons are v2, out of POC scope (§17).
+- **Fallback text is content, not error handling** (§13). Every AI-dependent
+  surface needs one before it ships.
+- Story skeletons are authored for now. AI-generated skeletons are v2 (§17).

content/origins/README.md

@@ -0,0 +1,26 @@
+# /content/origins — starting-state seeds
+
+Thin authored files, one per starting point. An origin is an **initial-state
+overlay** on the static shared world (`/content/world`): where the player begins,
+the opening situation, and seeds for dispositions, inventory, starting quest, and
+character-creation constraints. Values and id references — never maps or personas.
+
+At new-game the player picks an origin; code constructs the runtime canon log
+from origin + world content + character creation. See the spec in
+[`/docs/canon-log.md`](../../docs) (design under
+[`/docs/superpowers/specs`](../../docs/superpowers/specs)).
+
+## Scope (§17)
+
+Schema supports **N** origins; the POC authors **one**. One world, one map, fixed
+NPCs — origins change only the player's starting point and situation. More
+replayability, flat authoring cost.
+
+## Fields (summary)
+
+`id` · `display_name` · `description` · `start_location_id` · `situation[]` ·
+`opening_facts[]` · `disposition_overrides{}` · `inventory_grants[]` ·
+`start_quest_id` · `build_constraints{}`
+
+`humiliations` are never seeded — they are earned in play (§7/§9). Validated
+against `origin.schema.json`; every id must resolve in `/content/world`.

content/world/README.md

@@ -0,0 +1,16 @@
+# /content/world — static, ID-referenced content
+
+The shared world. Authored once, identical every playthrough. Everything here
+carries a stable string **id** that origins and the canon log reference. Never
+seed initial *state* here — that is an origin's job (`/content/origins`).
+
+```
+/locations   Towns, dungeons, points of interest. id → location
+/npcs        Persona + knowledge lists (charter §6). id → npc
+/quests      Story skeletons and quest definitions (charter §17). id → quest
+/items       Gear, consumables, cursed items (charter §7). id → item
+```
+
+New-game construction resolves every id an origin references (start location,
+start quest, granted items, seeded companions) against this content, and fails
+loudly if one is missing. Keep ids stable — a rename is a breaking change.

content/world/items/README.md

@@ -0,0 +1,9 @@
+# /content/world/items
+
+Item definitions — gear, consumables, cursed items — each with a stable **id**.
+An origin's `inventory_grants` resolves item ids here.
+
+Cursed/blessed items move Luck (charter §7): *a cursed blade granting +4 STR and
+−5 LCK is the most interesting item in this game.* The item system supports the
+STR/LCK split on day one. Numeric effects live in game state; only narrative-worthy
+items surface as canon-log facts.

content/world/locations/README.md

@@ -0,0 +1,8 @@
+# /content/world/locations
+
+The map — towns, dungeons, points of interest, each with a stable **id**. An
+origin's `start_location_id` resolves here; the canon log's `location` mirrors
+the current one (id + display name).
+
+POC scope (§17): one town, one dungeon (three fights, one boss). One reusable
+world; origins vary only where the player begins within it.

docker-compose.yml

@@ -0,0 +1,17 @@
+# Local dev — run the proxy in Docker with live reload.
+#   docker compose up --build
+# Ollama runs on the homelab (charter §4), not here — point OLLAMA_BASE_URL at it.
+services:
+  api:
+    build: ./api
+    ports:
+      - "8000:8000"
+    env_file:
+      - ./api/.env
+    environment:
+      PORT: 8000
+    # Mount source + reload so edits are live without a rebuild.
+    volumes:
+      - ./api/app:/app/app:ro
+    command: >
+      uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

docs/superpowers/specs/2026-07-09-canon-log-schema-design.md

@@ -0,0 +1,263 @@
+# Canon Log Schema & Origin Seeds — Design
+
+**Date:** 2026-07-09
+**Status:** approved (design); implementation plan to follow
+**Charter refs:** §2 (code owns state), §7 (Luck), §9 (companions/humiliations), §11 (canon log), §14 (cost/latency), §16 (layout)
+
+## Problem
+
+The canon log (§11) is the compact structured fact list code maintains and injects
+into every AI call. It is the spine that stops the AI contradicting itself, and
+§11 warns it is "nearly impossible to retrofit." It must be designed before any
+call site depends on it.
+
+Concurrently, the player should be able to select one of several **starting
+origins**, each with its own starting location and situation — for replayability
+without multiplying authoring work.
+
+## Reframe: origin = initial-state seed over a static world
+
+Rather than "multiple maps," an origin is an **initial-state overlay** on a single
+shared, static world. The world (one map, its towns, their fixed NPCs and
+personalities/knowledge) is authored once. An origin only seeds *where the player
+starts and the situation they are in*. This keeps charter §17 scope intact (one
+world authored) while raising replayability (N thin origin seeds).
+
+Confirmed scope: **schema supports N origins; the POC authors one.** An origin
+seeds: starting location, opening situation, initial dispositions, initial
+inventory/items, initial quest/objective, and player build constraints.
+
+## Architecture — three layers, one direction of flow
+
+```
+WORLD CONTENT (static)   /content/world      map · npcs(+knowledge) · quests · items
+        │ referenced by stable string id
+ORIGIN SEED (thin, N)    /content/origins    location + situation + dispo/quest/item/build seeds
+        │ new-game construct (origin + world + character creation)
+CANON LOG (runtime)      code-owned, saved   the §11 structure, injected into every AI call
+```
+
+Content and seed are **immutable inputs**; the canon log is the **single mutable
+product**. Nothing writes back up. This is §2 (code owns state) made concrete.
+
+### The in/out boundary
+
+What the canon log carries vs. what it must never carry:
+
+| In the log (narrative context → AI) | Not in the log |
+|---|---|
+| Player name, class, **luck _descriptor_** | Numeric Luck value — §7: AI must never be able to calculate it |
+| Current location (id + name) | Numeric stats, HP/MP, combat state — combat engine owns these |
+| Party dispositions | Full inventory contents — game state; only narrative items become facts |
+| Rolling recent events (≤5) | NPC knowledge lists — §6, injected only into that NPC's call |
+| Established facts (harvested `[FACT]`) | Available moves — §6, injected only into the NPC call |
+| Active quests | |
+| Humiliations (Banter, §9) | |
+
+Rationale: §7 requires the player to *feel* cursed but never *calculate* it, so the
+numeric Luck physically cannot enter the AI payload — only `luck_descriptor` does.
+Keeping stats/HP/inventory out keeps the payload small (§14) and separates combat
+state from narrative context.
+
+## Canon log schema (runtime state)
+
+```json
+{
+  "schema_version": 1,
+  "player": {
+    "name": "Aldric",
+    "class_id": "sellsword",
+    "luck_descriptor": "Fortune spits on you"
+  },
+  "location": { "id": "greywater_docks", "name": "the Greywater docks" },
+  "party": [
+    { "id": "brannoc_thane", "name": "Brannoc Thane", "disposition": 40 },
+    { "id": "cadwyn_vell",   "name": "Cadwyn Vell",   "disposition": 15 }
+  ],
+  "recent_events": [
+    "Arrived at Greywater by barge before dawn",
+    "Brannoc recognised the harbourmaster and went quiet"
+  ],
+  "established_facts": [
+    "the eastern bridge out of Greywater is washed out",
+    "the harbourmaster is named Oda Fenn"
+  ],
+  "active_quests": [
+    { "id": "find_the_ledger", "name": "The Missing Ledger",
+      "status": "active", "objective": "Find who took Fenn's ledger" }
+  ],
+  "humiliations": [
+    { "id": "h_0001", "text": "vomited on a shrine step in front of a priest",
+      "weight": 7, "turn": 3 }
+  ]
+}
+```
+
+| Field | Type | Notes |
+|---|---|---|
+| `schema_version` | int | Migration handle; saved logs upgraded on load. |
+| `player.name` | string | From character creation. |
+| `player.class_id` | enum | `sellsword` \| `assassin` \| `priest` (POC, §8). Id, not display name. |
+| `player.luck_descriptor` | string | **Only** Luck representation in the log (§7). Recomputed when numeric Luck drifts. Never the number. |
+| `location` | `{id, name}` | id → world content; name → prose. |
+| `party[]` | `{id, name, disposition}` | Companions. `disposition` int −100..100 (§9). |
+| `recent_events[]` | string[] | Rolling window, **cap 5** (§11). Code writes each terse line, drops oldest. |
+| `established_facts[]` | string[] | Durable. Harvested from `[FACT: …]` (§11) + origin seed. Deduped. Uncapped. |
+| `active_quests[]` | `{id, name, status, objective}` | `status`: `active` \| `complete` \| `failed`. Definitions in world content. |
+| `humiliations[]` | `{id, text, weight, turn}` | Banter memory (§9). **Append-only, stacks.** `weight` 1–10; `turn` drives decay of reference frequency. |
+
+Decisions:
+- Every entity carries both `id` (code keys/validates) and `name` (prompt reads) —
+  denormalized on purpose so the language model gets natural nouns and code gets
+  stable keys.
+- `recent_events` cap = 5 (§11 says "3–5"; ceiling taken). Hard rolling window.
+- `established_facts` uncapped — dropping a fact reintroduces the exact
+  contradiction §11 exists to prevent. If size bites (§14), fix is summarisation,
+  not truncation.
+- No global `npc_dispositions`; an NPC's live disposition rides only in its
+  `/npc/speak` call (§6). Add later if the Narrator needs town-wide stances.
+
+## Origin seed schema
+
+Lives in `/content/origins/<id>.json`. All heavy nouns are id references into
+world content; the origin owns only starting values.
+
+```json
+{
+  "schema_version": 1,
+  "id": "deserter",
+  "display_name": "The Deserter",
+  "description": "You walked away from a company that doesn't allow walking away.",
+  "start_location_id": "greywater_docks",
+  "situation": [
+    "Arrived at Greywater by barge before dawn, hood up",
+    "Down to your last coin and owed a favour you can't repay"
+  ],
+  "opening_facts": [
+    "the player deserted the Iron Kettle mercenary company",
+    "a bounty notice for the player circulates in the northern towns"
+  ],
+  "disposition_overrides": { "brannoc_thane": 40, "cadwyn_vell": 15 },
+  "inventory_grants": [
+    { "item_id": "worn_shortsword", "qty": 1 },
+    { "item_id": "coin", "qty": 3 }
+  ],
+  "start_quest_id": "find_the_ledger",
+  "build_constraints": {
+    "allowed_classes": ["sellsword", "assassin", "priest"],
+    "luck_modifier": 0
+  }
+}
+```
+
+| Field | Feeds | Notes |
+|---|---|---|
+| `id` / `display_name` / `description` | Origin-select screen | `description` is authored flavour shown at pick time. |
+| `start_location_id` | `location` | Must resolve to a world location, else new-game fails validation. |
+| `situation[]` | `recent_events` | Opening framing — why you're here, now. |
+| `opening_facts[]` | `established_facts` | What is already canon in this origin. |
+| `disposition_overrides{}` | `party[].disposition` (companions) / game-state NPC store (world NPCs) | Absolute starting values; omitted → world default. Companion ids land in the log; world-NPC ids land in game state and surface only at that NPC's `/npc/speak` call (§6) — the log has no global NPC-disposition field. |
+| `inventory_grants[]` | game-state inventory | Not the log — inventory is state (§2). Narrative items may also seed a fact. |
+| `start_quest_id` | `active_quests` | Nullable; resolved against world quest defs. `null` → no active quest. |
+| `build_constraints{}` | character creation | `allowed_classes` gates the picker; `luck_modifier` feeds Luck gen. Acts *before* the log exists. |
+
+`humiliations` is never seeded — earned in play only (§7/§9).
+
+## New-game construction
+
+```
+1. Player picks origin
+     load /content/origins/<id>.json → validate every id resolves in world content
+2. Apply build_constraints → character creation
+     class picker limited to allowed_classes
+     Luck generated per §7 (5-roll average) + luck_modifier
+3. Player creates character (name, class, rolled stats + Luck)
+4. Construct initial canon log:
+     player            ← creation (name, class_id) + luck→descriptor
+     location          ← start_location_id (+ name from world)
+     party             ← world default roster, companion dispositions ← disposition_overrides
+     (world-NPC dispositions from disposition_overrides → game-state NPC store, not the log)
+     recent_events     ← situation[]
+     established_facts ← opening_facts[]  (+ world always-true facts, if any)
+     active_quests     ← resolve(start_quest_id)   (or [])
+     humiliations      ← []   (always empty)
+     schema_version    ← 1
+     inventory_grants  → applied to game-state inventory (NOT the log)
+5. Log handed to the game loop; first Narrator call fires with it.
+```
+
+Construction reads all three layers and writes only the log. `build_constraints`
+firing at step 2 (before the log exists at step 4) is why an origin is a distinct
+schema, not a partial canon log: Luck is rolled during creation with the origin's
+modifier, and only its *descriptor* reaches the log. Validation is front-loaded at
+step 1 — a broken origin fails at new-game, loudly, not three scenes later.
+
+## Maintenance — turn-to-turn evolution
+
+Only code mutates the log. AI prose is never the source of truth (§2/§11); a
+`[FACT]` tag is a *request* to record, and code records the canonical string that
+later calls read.
+
+| Trigger | Mutation | Rule |
+|---|---|---|
+| Any role emits `[FACT: …]` | Append to `established_facts` (dedup) | §11 harvest; only way facts enter. |
+| A narrative beat resolves | Append terse line to `recent_events`, drop oldest past 5 | Code writes the line, not the AI. |
+| Validated `[ADJUST_DISPOSITION: ±n]` (§6) / approval change (§9) | Update `party[].disposition`, clamp −100..100 | Only after the move passes validation; invalid dropped. |
+| Event with embarrassment weight ≥ threshold | Append `{id, text, weight, turn}` to `humiliations` | Append-only, stacks forever (§9). |
+| Numeric Luck drifts (§7) | Recompute `player.luck_descriptor` | Number stays in game state; only descriptor mirrored. |
+| Quest state changes | Update `active_quests[].status` / objective | Definitions in world content; log tracks live status. |
+| Player moves | Update `location` | id + name from world content. |
+
+### Per-role injection
+
+The canon log is the shared base of every call; roles add role-specific extras.
+The log never grows per-role fields.
+
+| Role | Log + … |
+|---|---|
+| Narrator | log only |
+| NPC | log + this NPC's persona, `knowledge[]`, `available_moves[]` (§6) |
+| Improviser | log + the whitelisted event-change set (§7 validator) |
+| Banter | log, with `humiliations` as primary source (§9) |
+| Adjudicator | log + the current legal action set |
+
+## Storage & format
+
+```
+/content/
+  origins/        origin seeds (authored) + (later) validated against origin.schema.json
+  world/          static, id-referenced content
+    locations/  npcs/  quests/  items/
+  fallback/       degraded-DM text (§13) — NOT world content, no ids resolved against it
+/docs/
+  canon-log.md            living contract: this schema, field tables, example
+  schemas/
+    canon-log.schema.json  JSON Schema — validated on both client and api
+    origin.schema.json     JSON Schema — origin seed
+save file → serialized canon log (JSON) + game state (Luck number, inventory, stats)
+```
+
+- **JSON everywhere.** The log crosses the HTTP boundary to `/api` and seeds
+  prompts. The client (GDScript) models it as typed objects and serialises; `/api`
+  (Python) validates against the same `canon-log.schema.json`. One schema, two
+  languages, no drift.
+- **The log is part of the save.** Load a save → same log → same injected context.
+  Pairs with §10's per-encounter seeding: reproducible AI context, not just
+  reproducible combat. §11 rejected "last N messages"; persisting the *structured
+  log* keeps the anti-amnesia guarantee across save/quit.
+- **Schemas live in `/docs/schemas/`** — cross-cutting contracts both sides consume
+  (§16 doc split). Prose spec beside them in `/docs/canon-log.md`.
+
+## Consequences / follow-ups
+
+- `fallback/` stays outside `world/` (category: authored prose, not id-referenced).
+- Implementation will create the two JSON Schemas, the construction routine, the
+  maintenance hooks, and one authored POC origin + minimal world content to
+  exercise new-game end-to-end.
+- Out of scope here: the AI role prompts themselves, combat state, and the save
+  system's non-log portions. This spec defines the data contract they all consume.
+
+## Open questions
+
+None blocking. Thresholds (humiliation `weight` cutoff, Banter decay curve) are
+tuning values, deferred to implementation.