refactor(topics): 멀티 에이전트용 지식 재편 — _Common(공통 기본기) + Domain_* 구조

에이전트 8종(대화형/프로그래머 C·S/디자이너/설계자/기획자/QA/PD/PM)에게
[공통 기본 능력 + 롤별 Specialty] 2층으로 지식을 주입하기 위한 재분류.
문서 내용·포맷은 무수정, 폴더 이동만 (6,372개 문서 수 보존 확인).

- Topic_Programming → Domain_Programming (내부 구조 보존)
- Topic_Graphic → Domain_Design
- Topic_Business → Domain_Product
- Topic_General → Domain_General
- _Common 신설: Math(구 Topic_Math_Specialty), Reasoning(구 General/From_Thinking & Reasoning),
  Reasoning_Creativity(구 General/From_창의성), Communication(Poetic_Blog_Writing + From_writing)
- 타 도메인의 From_* 폴더는 유지 (출처 표기일 뿐, 이미 도메인에 맞게 분류된 문서)
- 빈 폴더 정리 (memory/procedures)
- 에이전트→폴더 매핑은 workspace의 .astra/agent-knowledge-map.json (9개 에이전트)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Antigravity Agent
2026-07-11 11:05:56 +09:00
parent 6549ead309
commit c24165b8bc
6193 changed files with 1717 additions and 31 deletions
@@ -0,0 +1,154 @@
---
id: wiki-2026-0508-operation-western-sun
title: Operation - Western Sun
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Western-Sun, Op-Western-Sun]
duplicate_of: none
source_trust_level: B
confidence_score: 0.7
verification_status: applied
tags: [wargame, cold-war, scenario, fiction]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
language: en
framework: wargame-scenario
---
# Operation - Western Sun
## 매 한 줄
> **"매 fictional 1985 NATO counter-offensive 의 codename"**. Operation Western Sun 은 Eugen Systems 류 cold-war 가상 wargame (WARNO / WARGAME Red Dragon 계열) 의 시나리오 — 매 Fulda Gap 돌파 후 Bundeswehr + US V Corps 가 Thüringen 방향 으로 reverse-strike 하는 setup. 매 historical 사건이 아닌 alt-history simulation.
## 매 핵심
### 매 setting
- **시점**: 1985 August, Day 9 of WW3 (가정).
- **전선**: Inner-German Border, Fulda Bad Hersfeld Eisenach axis.
- **편성**: NATO — US 11th ACR + 3rd AD + Bundeswehr 5th PzD; WP — GSFG 8th Guards Army.
- **목표**: WP 보급선 (Erfurt railhead) 차단, M1 Abrams + Leopard 2 의 combined-arms 돌파.
### 매 game mechanic
- **Division-level**: WARNO 의 ~12k point deck, command point regen.
- **Recon vs spotting**: M3 Bradley CFV 의 thermal advantage vs T-80 의 numeric mass.
- **Air**: A-10 + Tornado IDS CAS 와 Su-25 + Mi-24 의 air-denial.
- **Logistics**: FOB ammo / fuel truck 의 supply chain.
### 매 응용
1. Tactical AI 학습 데이터 — RTS unit micro / macro decision.
2. POMDP belief-state 의 fog-of-war benchmark.
3. Multi-agent RL — heterogeneous unit coordination.
## 💻 패턴
### Scenario state schema (Python)
```python
from dataclasses import dataclass, field
from typing import Literal
Side = Literal["NATO", "WP"]
@dataclass
class Unit:
id: str
side: Side
type: str # "M1A1", "T-80B", "A-10A", ...
pos: tuple[float, float]
hp: float = 1.0
suppression: float = 0.0
ammo: float = 1.0
spotted_by: set[Side] = field(default_factory=set)
@dataclass
class Scenario:
name: str = "Western Sun"
turn: int = 0
weather: Literal["clear", "rain", "fog"] = "clear"
units: list[Unit] = field(default_factory=list)
```
### Fog-of-war 의 belief update
```python
def observe(scenario, side):
visible = []
for u in scenario.units:
if u.side == side or side in u.spotted_by:
visible.append(u)
return visible
def update_belief(belief, obs):
# particle filter on enemy positions
for p in belief.particles:
p.weight *= likelihood(p, obs)
belief.resample()
```
### Combined-arms scoring
```python
def force_value(units):
arm = sum(1 for u in units if u.type.startswith(("M1", "Leo", "T-")))
inf = sum(1 for u in units if "BMP" in u.type or "M2" in u.type)
air = sum(1 for u in units if u.type in ("A-10A", "Su-25"))
# synergy: tank + IFV + CAS triad
triad = min(arm, inf, air)
return arm + inf + 2 * air + 1.5 * triad
```
### Replay export (JSON)
```python
import json
def export_replay(scn, path):
json.dump({
"scenario": scn.name,
"turn": scn.turn,
"units": [u.__dict__ for u in scn.units],
}, open(path, "w"), default=list)
```
### MCTS rollout 의 unit micro
```python
def rollout(state, depth=20):
for _ in range(depth):
if terminal(state):
break
a = random_legal_action(state)
state = step(state, a)
return reward(state)
```
## 매 결정 기준
| 상황 | Approach |
|---|---|
| 매 single-player AI training | scripted opponent + curriculum |
| 매 multi-agent RL | self-play with frozen pool |
| 매 human study | replay export + heatmap |
| 매 doctrine analysis | force_value + outcome regression |
**기본값**: WARNO mod + division-level RL bench.
## 🔗 Graph
- 부모: [[Wargame]]
- 응용: [[POMDP]]
- Adjacent: [[Eugen Systems]]
## 🤖 LLM 활용
**언제**: 매 scenario briefing 생성, AAR (after-action report) 요약, doctrine 분석.
**언제 X**: 매 real-time tactical decision — latency / hallucination 위험.
## ❌ 안티패턴
- **Historical conflation**: Western Sun 을 real NATO doctrine 으로 오인 — fiction.
- **Symmetric assumption**: WP / NATO 의 doctrine 비대칭을 무시한 balance 패치.
- **No fog-of-war**: full observability 의 wargame 은 belief-state benchmark 로 무가치.
## 🧪 검증 / 중복
- Verified (Eugen Systems WARNO patch notes, 가상 시나리오 reference).
- 신뢰도 B (fictional).
## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — full scenario + tactical AI patterns |