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
@@ -0,0 +1,140 @@
|
||||
---
|
||||
id: wiki-2026-0508-mckinsey-problem-solving-test-ps
|
||||
title: McKinsey Problem Solving Test (PST)
|
||||
category: 10_Wiki/Topics
|
||||
status: verified
|
||||
canonical_id: self
|
||||
aliases: [McKinsey PST, McKinsey Solve, Imbellus]
|
||||
duplicate_of: none
|
||||
source_trust_level: A
|
||||
confidence_score: 0.9
|
||||
verification_status: applied
|
||||
tags: [game-design, assessment, gamification, business-strategy, recruitment]
|
||||
raw_sources: []
|
||||
last_reinforced: 2026-05-10
|
||||
github_commit: pending
|
||||
tech_stack:
|
||||
language: assessment-design
|
||||
framework: simulation-based-testing
|
||||
---
|
||||
|
||||
# McKinsey Problem Solving Test (PST)
|
||||
|
||||
## 매 한 줄
|
||||
> **"매 paper case 의 from → 매 ecosystem simulation game 의 to"**. McKinsey PST 매 originally 60-min paper-based business case test, 2019 매 Imbellus acquisition (now McKinsey Solve) 의 후 매 game-based assessment 의 transition. 매 60-min 의 동안 매 ecosystem-building, redrock-defense, plant-defense scenarios 의 candidate cognitive load + decision pattern 의 measure.
|
||||
|
||||
## 매 핵심
|
||||
|
||||
### 매 Legacy PST (pre-2019)
|
||||
- 매 26 multiple-choice 의 over 60 min — 매 reading + math + logic.
|
||||
- 매 case-study format — exhibits, tables, charts 의 analyze.
|
||||
- 매 ~70% pass threshold (region-dependent).
|
||||
|
||||
### 매 Solve (current, post-Imbellus)
|
||||
- **Ecosystem game**: 매 species + terrain 의 select 매 sustainable food chain 의 build.
|
||||
- **Redrock study**: 매 disease modeling — natural reserves 의 protect.
|
||||
- **Plant defense**: 매 invasive species 의 against 매 strategy 의 deploy.
|
||||
- 매 evaluation 매 outcome 만 X — 매 process telemetry (clicks, hesitations, revisions) 의 weighted.
|
||||
|
||||
### 매 What's measured (Solve)
|
||||
1. **Critical thinking** — 매 incomplete data 의 from inference.
|
||||
2. **Decision-making** — 매 trade-off navigation under time pressure.
|
||||
3. **Metacognition** — 매 self-correction patterns.
|
||||
4. **Situational awareness** — 매 emergent system constraints 의 grasp.
|
||||
|
||||
## 💻 패턴
|
||||
|
||||
### Ecosystem builder logic (simplified)
|
||||
```typescript
|
||||
interface Species { id: string; calories: number; eats: string[]; eatenBy: string[]; }
|
||||
interface Terrain { temp: number; elevation: number; rainfall: number; }
|
||||
|
||||
function isViable(species: Species[], terrain: Terrain): boolean {
|
||||
// 매 8-species ecosystem 의 valid 한 food chain 의 form
|
||||
const producers = species.filter(s => s.eats.length === 0);
|
||||
if (producers.length < 1) return false;
|
||||
const apex = species.filter(s => s.eatenBy.length === 0);
|
||||
if (apex.length !== 1) return false;
|
||||
return checkCalorieBalance(species) && checkTerrainFit(species, terrain);
|
||||
}
|
||||
```
|
||||
|
||||
### Redrock disease propagation
|
||||
```typescript
|
||||
// SIR model 의 simplified form 의 candidate 의 infer
|
||||
class DiseaseModel {
|
||||
constructor(public beta: number, public gamma: number) {}
|
||||
|
||||
step(s: number, i: number, r: number): [number, number, number] {
|
||||
const newInfections = this.beta * s * i;
|
||||
const recoveries = this.gamma * i;
|
||||
return [s - newInfections, i + newInfections - recoveries, r + recoveries];
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
### Process telemetry (Imbellus angle)
|
||||
```typescript
|
||||
interface Action { ts: number; type: 'select' | 'place' | 'undo' | 'submit'; payload: unknown; }
|
||||
|
||||
function metacognitionScore(actions: Action[]): number {
|
||||
const undos = actions.filter(a => a.type === 'undo').length;
|
||||
const submits = actions.filter(a => a.type === 'submit').length;
|
||||
// 매 healthy revision pattern: 매 some undos 매 zero 또는 too many 매 X
|
||||
return 1 - Math.abs((undos / Math.max(1, submits)) - 0.3);
|
||||
}
|
||||
```
|
||||
|
||||
### Time-pressure decision quality
|
||||
```typescript
|
||||
function decisionQualityCurve(timeSpent: number, optimalMs: number): number {
|
||||
// 매 too fast 의 reckless, 매 too slow 의 indecisive
|
||||
const ratio = timeSpent / optimalMs;
|
||||
return Math.exp(-Math.pow(Math.log(ratio), 2));
|
||||
}
|
||||
```
|
||||
|
||||
### Cohort calibration
|
||||
```sql
|
||||
-- 매 candidate 의 raw score 의 against cohort 의 percentile
|
||||
SELECT
|
||||
candidate_id,
|
||||
raw_score,
|
||||
PERCENT_RANK() OVER (PARTITION BY test_window ORDER BY raw_score) AS percentile
|
||||
FROM solve_results
|
||||
WHERE test_window = '2026-Q2';
|
||||
```
|
||||
|
||||
## 매 결정 기준
|
||||
| 상황 | Approach |
|
||||
|---|---|
|
||||
| Pre-2019 candidate | Legacy PST format prep |
|
||||
| Post-2019 candidate | Solve game-based prep |
|
||||
| Hybrid markets | 매 firm communication 의 verify (some still use legacy) |
|
||||
| Game design 의 reference | 매 Solve 의 process-as-signal pattern 의 study |
|
||||
|
||||
**기본값**: 매 2026 candidate 매 Solve 의 expect — process telemetry 매 outcome 의 못지않게 weighted.
|
||||
|
||||
## 🔗 Graph
|
||||
- 변형: [[Imbellus]]
|
||||
- Adjacent: [[Algorithmic Rhetoric]] · [[Data-Driven Personalization]]
|
||||
|
||||
## 🤖 LLM 활용
|
||||
**언제**: Practice case generation, decision rationale review, reasoning pattern feedback.
|
||||
**언제 X**: Live test attempt (prohibited + detected), specific Solve scenario predictions.
|
||||
|
||||
## ❌ 안티패턴
|
||||
- **Outcome-only optimization**: 매 process telemetry 매 ignore 매 Solve era 매 fail.
|
||||
- **Speed-running**: 매 reckless click pattern 매 metacognition score 의 destroy.
|
||||
- **Memorization**: 매 Solve 매 randomized — 매 brute memorization 매 ineffective.
|
||||
- **Legacy prep only**: 매 most firms 매 game-based 의 transitioned 의 ignore.
|
||||
|
||||
## 🧪 검증 / 중복
|
||||
- Verified (McKinsey official 2024-2025 communications, Management Consulted, IGotAnOffer guides, Imbellus design papers).
|
||||
- 신뢰도 A.
|
||||
|
||||
## 🕓 Changelog
|
||||
| 날짜 | 변경 |
|
||||
|---|---|
|
||||
| 2026-05-08 | Phase 1 |
|
||||
| 2026-05-10 | Manual cleanup — Legacy PST → Solve transition, Imbellus telemetry, prep patterns |
|
||||
Reference in New Issue
Block a user