c24165b8bc
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>
5.1 KiB
5.1 KiB
id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack
| id | title | category | status | canonical_id | aliases | duplicate_of | source_trust_level | confidence_score | verification_status | tags | raw_sources | last_reinforced | github_commit | tech_stack | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| wiki-2026-0508-5r-structure | 5R Structure | 10_Wiki/Topics | verified | self |
|
none | B | 0.8 | applied |
|
2026-05-10 | pending |
|
5R Structure
매 한 줄
"매 5R 은 매 communication / project / data lifecycle 을 매 5 phase 로 grouping 하는 mnemonic". 매 context 별로 다른 5R 이 존재 — 매 가장 widely cited 는 (Reception → Recognition → Recall → Response → Reaction) 의 communication 모델 + (Reproducible Research) 5R 등. 매 2026 에서 매 LLM agent design (RAG → 매 5R 변형) 에도 적용된다.
매 핵심
매 가장 흔한 5R 변형
- Communication 5R (Schramm 후속): Reception, Recognition, Recall, Response, Reaction.
- Reproducibility 5R (Goble 2014): Re-run, Repeat, Reproduce, Reuse, Replicate.
- Waste hierarchy 5R: Refuse, Reduce, Reuse, Repurpose, Recycle.
- Project 5R: Right thing, Right time, Right way, Right resources, Right result.
- Customer-relationship 5R: Reach, Relate, Retain, Reward, Refer.
매 공통 구조
- 매 sequential 또는 매 layered.
- 매 mnemonic 이 핵심 — 매 R 알파벳 시작 단어 force.
- 매 framework, not theory — 매 prescriptive checklist.
매 응용
- Research project: 매 reproducibility 5R 로 매 paper review.
- Customer onboarding: 매 Reach→Refer 5R.
- AI agent design: 매 RAG (Retrieve, Rerank, Read, Reason, Respond) — 매 modern 5R.
💻 패턴
Pattern 1: Reproducibility 5R 검증
# Goble 2014 — 5R reproducibility checklist
checklist = {
"rerun": lambda paper: paper.has("docker_image"),
"repeat": lambda paper: paper.has("seeds_fixed"),
"reproduce": lambda paper: paper.has("data_open"),
"reuse": lambda paper: paper.has("license_permissive"),
"replicate": lambda paper: paper.has("methods_section_complete"),
}
score = sum(fn(paper) for fn in checklist.values()) / 5
Pattern 2: 매 RAG 5R agent
def rag_5r(query: str, kb) -> str:
docs = retrieve(query, kb, k=20) # Retrieve
docs = rerank(query, docs, k=5) # Rerank
chunks = read_full(docs) # Read
plan = reason(query, chunks) # Reason
return respond(plan) # Respond
Pattern 3: 매 Customer 5R funnel
SELECT stage,
COUNT(*) AS users,
LAG(COUNT(*)) OVER (ORDER BY stage_order) AS prev,
COUNT(*)::float / NULLIF(LAG(COUNT(*)) OVER (ORDER BY stage_order), 0) AS conv
FROM (
SELECT user_id, 'reach' AS stage, 1 AS stage_order FROM impressions
UNION ALL SELECT user_id, 'relate', 2 FROM signups
UNION ALL SELECT user_id, 'retain', 3 FROM active_30d
UNION ALL SELECT user_id, 'reward', 4 FROM rewards_claimed
UNION ALL SELECT user_id, 'refer', 5 FROM referrals
)
GROUP BY stage, stage_order
ORDER BY stage_order;
Pattern 4: Waste 5R audit
items = [...] # household / office items
for item in items:
if can_refuse(item): act = "refuse"
elif can_reduce(item): act = "reduce"
elif can_reuse(item): act = "reuse"
elif can_repurpose(item):act = "repurpose"
else: act = "recycle"
Pattern 5: 매 Communication 5R review
- Reception: was the message received? (delivery confirmation)
- Recognition: was sender / topic identified?
- Recall: can the receiver recall key points 24h later?
- Response: did receiver respond?
- Reaction: was behavior changed?
매 결정 기준
| 상황 | Which 5R |
|---|---|
| 매 research paper review | Reproducibility 5R |
| 매 sustainability audit | Waste 5R |
| 매 customer growth | Reach→Refer |
| 매 LLM agent | RAG 5R (Retrieve/Rerank/Read/Reason/Respond) |
| 매 communication training | Reception→Reaction |
기본값: 매 context 명시 — 매 "5R" 단독 의 ambiguous. 매 always qualify (예: "RAG 5R", "Waste 5R").
🔗 Graph
- 부모: Pyramid Principle · MECE + Pyramid Principle--
- 변형: Rule of Three
- 응용: Knowledge synthesis · Process_Reflection_Template
- Adjacent: Working-Backwards · Outside-Thinking
🤖 LLM 활용
언제: 매 structured output template 생성 — 매 5-bullet checklist agent. 매 mnemonic 이 매 LLM recall 에 유리. 언제 X: 매 단순 list — 매 5R force-fit 의 contrived.
❌ 안티패턴
- Force-fit: 매 4 또는 6 step 인데 매 5R 강제 → 매 awkward bucket.
- Buzzword usage: 매 "5R framework" 만 언급, 매 actual 5 step 의 unclear.
- Cross-domain confusion: 매 RAG 5R + Waste 5R 동일시 — 매 unrelated.
🧪 검증 / 중복
- Verified (Goble Better Software Practices 2014, Schramm communication model derivatives).
- 신뢰도 B+.
🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — 5 variants + RAG 5R modern application |