MCP HubMCP Hub
스킬 목록으로 돌아가기

integrate-gestalt

pjt222
업데이트됨 2 days ago
3 조회
17
2
17
GitHub에서 보기
테스팅aitesting

정보

이 스킬은 `expand-awareness`에서 얻은 다양한 도메인의 원시 인식을 종합하여 단일한 창발적 통찰로 만듭니다. 각 도메인 간의 긴장과 공명을 매핑하여 조기에 결론을 내리지 않으면서도 일관된 "게슈탈트" 전체를 형성합니다. `expand-awareness` 이후, `express-insight` 이전에 사용하여 단일 관점으로는 도출할 수 없는 통합된 결론을 표현하세요.

빠른 설치

Claude Code

추천
기본
npx skills add pjt222/agent-almanac -a claude-code
플러그인 명령대체
/plugin add https://github.com/pjt222/agent-almanac
Git 클론대체
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/integrate-gestalt

Claude Code에서 이 명령을 복사하여 붙여넣어 스킬을 설치하세요

문서

Integrate Gestalt

Form a coherent whole from the panoramic perception produced by expand-awareness — not by averaging, compromising, or selecting the best domain's answer, but by identifying the emergent pattern that could not have arisen from any single perspective alone.

When to Use

  • expand-awareness has surfaced raw perception from multiple domains and the observations need to become a unified insight
  • Multiple domain perspectives are available but no single one accounts for all the evidence
  • A problem has been analyzed from several angles and the separate analyses need to become more than a list
  • The question "what does this all mean, taken together?" has no obvious answer
  • When synthesis keeps collapsing into "pick the best domain" rather than forming something new
  • Before express-insight, which requires a formed gestalt as its input

Inputs

  • Required: Multi-domain observations from expand-awareness (or equivalent panoramic perception)
  • Optional: The original question or problem that prompted the multi-domain scan
  • Optional: Known constraints that the gestalt must satisfy
  • Optional: Prior failed integration attempts (what collapsed into single-domain answers)

Procedure

Step 1: Map Tensions

For each pair of domains identified in the panoramic perception, characterize how they relate. The three possible relationships are tension (they disagree), resonance (they reinforce from different angles), and orthogonality (they address unrelated aspects).

Use the tension-resonance map:

Tension-Resonance Map
+-------------------+-------------------+-------------------------------+
| Domain Pair       | Relationship      | Detail                        |
+-------------------+-------------------+-------------------------------+
| A vs B            | tension /         |                               |
|                   | resonance /       |                               |
|                   | orthogonal        |                               |
|   Evidence:       |                   | What specifically disagrees,  |
|                   |                   | reinforces, or is unrelated?  |
|   Implication:    |                   | What does this relationship   |
|                   |                   | suggest for the whole?        |
+-------------------+-------------------+-------------------------------+
| A vs C            | ...               | ...                           |
+-------------------+-------------------+-------------------------------+
| B vs C            | ...               | ...                           |
+-------------------+-------------------+-------------------------------+

Fill one row for every domain pair. For N domains there are N(N-1)/2 pairs. If this exceeds ten rows, group related domains first and map between groups.

Prioritize tensions — they carry the most integrative information. Resonances confirm; orthogonalities can be set aside; tensions demand resolution, and the gestalt is found in how they resolve.

Got: A completed tension-resonance map where every domain pair has a characterized relationship with specific evidence. At least one genuine tension is identified — if there are no tensions, the domains may not be different enough to produce emergence.

If fail: If all pairs show resonance, the domains are agreeing at a surface level. Dig deeper: where do they agree for different reasons? Agreement-for-different-reasons is a hidden tension. If no relationships can be characterized, the panoramic perception from expand-awareness may be too shallow — return and deepen the domain-specific observations before attempting integration.

Step 2: Find the Figure

In Gestalt psychology, the figure emerges from the ground. The ground is the tension-resonance map from Step 1. The figure is the dominant pattern that unifies the most domains with the fewest contradictions.

  1. Scan the map for clusters: which groups of domains resonate with each other? These clusters suggest candidate figures
  2. For each candidate figure, ask: "What single perspective makes sense of the most observations?"
  3. The figure is not a compromise (weakening each domain until they agree) nor a selection (choosing the strongest domain). It is a new frame that recontextualizes the domain observations
  4. Test: state the candidate figure in one sentence. Does it feel like it belongs to one of the input domains? If yes, it is not yet a gestalt — it is a domain answer wearing a disguise
  5. Look specifically at the tensions: the true figure often lives in the space between the disagreeing domains, not in either domain's position

Signs the figure is emerging:

  • Multiple tensions resolve simultaneously under the same reframe
  • Domain observations that seemed contradictory become complementary aspects of the same phenomenon
  • The figure explains why each domain saw what it saw, including why they disagreed

Got: One or two candidate figures articulated as single sentences. Each candidate recontextualizes the domain observations rather than selecting among them. The candidate accounts for at least the major tensions in the map.

If fail: If no figure emerges, the integration may be premature. Two recovery paths: (a) return to expand-awareness and add a domain that was missing — sometimes the figure cannot form because a key perspective is absent; (b) sit with the tensions without forcing resolution — some gestalts need incubation rather than effort. Note the current state and return later.

Step 3: Test the Whole

The candidate gestalt from Step 2 must survive three tests before it is accepted.

Test A — Tension accounting: Walk through every tension from Step 1. Does the gestalt resolve it, reframe it, or explicitly acknowledge it as an irreducible trade-off? Unaddressed tensions indicate a premature gestalt.

Test B — Single-domain origin: Could this insight have come from within a single domain? If a domain specialist would nod and say "yes, we already knew that," the gestalt has collapsed back into a domain answer. A true gestalt surprises every domain — each recognizes its contribution but not the whole.

Test C — Coherence under rotation: Mentally approach the gestalt from each domain's perspective in turn. Does it hold its shape, or does it look different depending on which domain you view it from? A robust gestalt is the same insight viewed from any angle; a fragile one changes meaning under rotation.

Scoring:

  • All three tests pass: proceed to Step 4
  • Test A fails: the gestalt is incomplete — return to Step 2 with the unresolved tensions as additional constraints
  • Test B fails: the gestalt is not emergent — return to Step 2 and explicitly exclude single-domain framings
  • Test C fails: the gestalt is not coherent — it may be two separate insights masquerading as one. Split and test each half independently

Got: The candidate gestalt passes all three tests, or the failure mode is clearly identified and guides a return to Step 2.

If fail: If the candidate fails repeatedly after multiple iterations, consider that the domains may not form a natural gestalt for this problem. Not every multi-domain observation produces emergence — sometimes the honest answer is a structured list of domain perspectives with their tensions mapped. Deliver the tension-resonance map as the output rather than forcing a false unity.

Step 4: Name the Insight

Articulate the gestalt in a single sentence that a domain specialist would not have written from within their domain alone. This sentence is the deliverable.

  1. Write the sentence. It should be:
    • Specific enough to be actionable or falsifiable
    • General enough to encompass all contributing domains
    • Surprising to at least two of the input domains
    • Free of jargon from any single domain (or using jargon deliberately recontextualized)
  2. Test the sentence against the three criteria from Step 3 one final time
  3. Optionally, add a one-paragraph expansion that traces how the gestalt emerged from the domain contributions — this is the provenance, not the insight itself
  4. Record which domains contributed, which tensions were key, and what the figure-ground relationship was — this metadata supports future integration attempts

The named insight, together with its provenance, becomes the input to express-insight for communication.

Got: A single sentence capturing the gestalt, accompanied by a brief provenance paragraph. The sentence passes the "no single domain" test. Reading it, a practitioner of any contributing domain recognizes their field's contribution but could not have arrived at the statement alone.

If fail: If the sentence keeps collapsing into domain language, try the negation test: state what the gestalt is NOT. "This is not a security recommendation, and not a performance optimization, and not an architectural pattern — it is [the gestalt]." The negations clear the domain frames and create space for the emergent formulation.

Validation

  • A tension-resonance map was completed for all domain pairs with specific evidence
  • At least one genuine tension (not a difference of emphasis) was identified
  • The candidate gestalt was articulated as a reframe, not a compromise or selection
  • Test A passed: all major tensions are resolved, reframed, or acknowledged
  • Test B passed: no single domain could have produced this insight alone
  • Test C passed: the gestalt holds its shape when viewed from each domain's perspective
  • The final insight is expressed in a single sentence with provenance

Pitfalls

  • Averaging: Weakening each domain's position until they superficially agree. This produces mush, not gestalt. If the integration feels bland, it is averaging
  • King-making: Selecting the strongest domain's answer and dressing it in multi-domain language. Test B catches this — if one domain specialist would nod unsurprised, it is king-making
  • Premature closure: Accepting the first candidate figure without testing it against tensions. The first figure that emerges is often the most obvious, not the most integrative
  • Forced unity: Insisting that a gestalt must exist when the domains are genuinely orthogonal. Orthogonal domains produce structured lists, not gestalts — and that is a valid outcome
  • Jargon blending: Combining technical terms from multiple domains into a sentence that sounds integrative but means nothing. Every term in the final sentence should be independently meaningful

Related Skills

  • expand-awareness — produces the raw panoramic perception that this skill integrates; always precedes integrate-gestalt
  • express-insight — communicates the formed gestalt to its audience; always follows integrate-gestalt
  • build-coherence — selects between competing options using structured evaluation; integrate-gestalt forms a new whole rather than choosing among existing options
  • brahma-bhaga — creates from void; integrate-gestalt creates from abundance (multiple filled perspectives)
  • meditate — clears prior context to enable clean perception; useful before expand-awareness, which precedes this skill
  • coordinate-reasoning — manages information flow in multi-path evaluation; complementary when the gestalt involves coordinating multiple reasoning threads

GitHub 저장소

pjt222/agent-almanac
경로: i18n/caveman-lite/skills/integrate-gestalt
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

연관 스킬

evaluating-llms-harness

테스팅

이 Claude Skill은 MMLU, GSM8K를 포함한 60개 이상의 표준화된 학술 과제에서 LLM 성능을 벤치마크하기 위해 lm-evaluation-harness를 실행합니다. 개발자들이 모델 품질을 비교하고, 학습 진행 상황을 추적하거나 학술 결과를 보고할 수 있도록 설계되었습니다. 이 도구는 HuggingFace와 vLLM 모델을 포함한 다양한 백엔드를 지원합니다.

스킬 보기

cloudflare-cron-triggers

테스팅

이 스킬은 cron 표현식을 사용하여 Worker를 스케줄링하기 위한 Cloudflare Cron Triggers 구현에 관한 포괄적인 지식을 제공합니다. 주기적 작업, 유지보수 작업, 자동화된 워크플로우 설정 방법을 다루며, 잘못된 cron 표현식이나 시간대 문제 같은 일반적인 이슈들을 해결하는 방법을 포함합니다. 개발자들은 이를 통해 스케줄된 핸들러 구성, cron 트리거 테스트, Workflows 및 Green Compute와의 연동 작업을 수행할 수 있습니다.

스킬 보기

webapp-testing

테스팅

이 Claude Skill은 Python 스크립트를 통해 로컬 웹 애플리케이션을 테스트하기 위한 Playwright 기반 툴킷을 제공합니다. 프론트엔드 검증, UI 디버깅, 스크린샷 캡처, 로그 확인 기능을 지원하며 서버 라이프사이클을 관리합니다. 브라우저 자동화 작업에 사용하되 컨텍스트 오염을 방지하기 위해 소스 코드를 읽지 않고 스크립트를 직접 실행하세요.

스킬 보기

finishing-a-development-branch

테스팅

이 스킬은 테스트 통과를 확인한 후 체계적인 통합 옵션을 제시하여 개발자가 완성된 작업을 마무리하도록 돕습니다. 구현이 완료된 후 머지, PR 생성, 브랜치 정리와 같은 워크플로우를 안내합니다. 코드가 준비되고 테스트가 완료되었을 때 개발 프로세스를 체계적으로 마무리하기 위해 사용하세요.

스킬 보기