사용 방법
- 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
- 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
- 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/yogsoth-ai/de-anthropocentric-research-engine --skill select-next-pair시작 전 확인
- 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
- 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서
원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.
select-next-pair
Purpose
Select the next comparison pair to maximize expected information gain while respecting unresolved uncertainty and coverage.
Input contract
yaml
required: [candidate_set, current_ratings, uncertainty_matrix, comparison_budget]
optional: [pair_history, exploration_policy, exclusion_constraints]
constraints: [candidate IDs are stable; expected information gain uses the declared model and available evidence]Procedure
- Enumerate eligible unjudged pairs and compute their current uncertainty and decision relevance.
- Estimate expected information gain under the declared comparison model, including the value of resolving a tie or cycle.
- Exclude pairs blocked by scope or data constraints, then select the highest-value eligible pair.
- Return the selected pair, rationale, alternatives considered, and the budget impact.
Output contract
yaml
produces: [selected_pair, information_gain_estimate, candidate_pair_table, selection_rationale]
delta_fields: [findings, evidence_updates, uncertainties, decisions, recommended_jumps]Quality gates
- The selected pair is eligible, not already resolved under the same model, and has a stated information-gain rationale.
- If all pairs are exhausted or blocked, return an explicit stop reason.
Failure and counterexamples
Do not select a pair solely because it is adjacent in a list. Expected information gain cannot be claimed when uncertainty inputs are absent.
Provenance map
- resolved: pair-selector
