SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsinsane-research-query

insane-research-query

Writing & research

View repository

Updated 10-15 21:46

About this skill

Insane Research Query Builder for Codex

Transform vague research ideas into structured, actionable research queries.

Read first:

  • $PLUGIN_ROOT/skills/insane-research-query/references/query_schema.json

This skill is used when the user wants to refine a vague idea into a research-ready brief before full-scale deep research.

Questioning principles (shared/questioning-policy.md §A·§1·§2c)

In Codex CLI, AskUserQuestion a multiple-choice card UI like does not exist. shared/questioning-policy.md §Aand is replaced with chat numbered blocks. Do not ask what can be inferred (§1), and if the request is already specific, generate the brief immediately without over-questioning (§2c). Match the language of all questions/options/deliverables to the user's input language.

Codex Workflow

Phase 1: Discovery (topic·type)

Only when the topic is unclear, ask once, using §A numbered blocks, only about the biggest unknown:

어떤 유형의 리서치인가요?
1. Exploratory — 무엇이 존재하는지 탐색, 지형 매핑
2. Comparative — 기술/접근/제품 비교
3. Analytical — 원인·효과·메커니즘 심층 분석
4. Predictive — 미래 트렌드·전망·예측
(번호 또는 문장으로 답해도 됩니다)

If the answer indicates they want to see examples, load and present them from the following: $PLUGIN_ROOT/skills/insane-research-main/examples/

Phase 2: Detailed Scoping (scope·source quality)

After establishing the core topic, confirm only the truly ambiguous constraints using §A numbered blocks (if inferable, proceed with defaults):

지리적 범위는?
1. Global (추천) — 전 세계 관점
2. US/North America — 미국·북미 중심
3. Asia-Pacific — APAC 중심 (한국 포함)
4. Europe — 유럽 시장 중심
(문장으로 직접 지정해도 됩니다)

필요한 소스 품질은?
1. B - High quality (추천) — 학술 + 공식 문서 + 검증된 리포트
2. A - Academic only — 피어리뷰 논문·메타분석만
3. C - Moderate — 전문가 의견·사례 연구 포함
4. D - Broad coverage — 프리프린트·전문가 블로그까지 폭넓게
(모르면 1번으로 진행하겠습니다)

Phase 3: Query Generation (query generation)

After gathering all inputs, generate:

  1. Structured JSON query — schema compliance: $PLUGIN_ROOT/skills/insane-research-query/references/query_schema.json
  2. Human-readable research brief — Markdown
  3. Execution checklist — for quality verification

JSON query structure

{
  "task": {
    "title": "[5-15 단어 간결한 제목]",
    "objective": "[명확한 리서치 목표 진술]",
    "type": "exploratory|comparative|analytical|predictive|evaluative"
  },
  "context": {
    "background": "[왜 이 리서치가 중요한가]",
    "audience": "technical|executive|academic|general|policy_maker",
    "use_case": "[리서치가 어떻게 쓰일지]",
    "prior_knowledge": ["가정 1", "가정 2"]
  },
  "questions": {
    "primary": "[메인 리서치 질문]",
    "secondary": ["서브 질문 1", "서브 질문 2", "서브 질문 3"],
    "hypotheses": ["검증 가능한 가정 1"],
    "exclusions": ["범위 밖 주제 1"]
  },
  "constraints": {
    "timeframe": {"start": "2024-01-01", "end": "present", "focus_period": "2025-2026"},
    "geography": {"scope": "global", "regions": [], "exclude_regions": []},
    "sources": {"required_types": ["peer_reviewed", "industry_reports"], "min_quality": "B", "language": ["en"]}
  },
  "output": {
    "format": "comprehensive_report",
    "length": {"min_words": 3000, "max_words": 10000},
    "structure": {"include_executive_summary": true, "include_bibliography": true, "generate_website": false},
    "citation_style": "APA",
    "tone": "professional"
  },
  "keywords": ["keyword1", "keyword2"],
  "special_instructions": []
}

Human-readable brief

# Research Brief: [Title]

## Objective
[명확한 진술]

## Research Questions
### Primary Question
> [메인 질문]

### Secondary Questions
1. [서브 질문 1]
2. [서브 질문 2]

## Scope & Constraints
| Dimension | Specification |
|-----------|--------------|
| Timeframe | [기간] |
| Geography | [범위] |
| Min Quality | Grade [X] |

## Execution Checklist
- [ ] Primary question fully answered
- [ ] All secondary questions addressed
- [ ] Sources meet quality threshold
- [ ] Citations properly formatted

Phase 4: Confirmation and Handoff

After showing the brief, confirm the next action using §A numbered blocks:

쿼리가 괜찮으면 바로 리서치를 시작할까요?
1. 지금 리서치 시작 (추천) — 이 쿼리로 즉시 딥리서치 실행
2. 쿼리만 저장 — JSON 쿼리를 파일로 저장해 두기
3. 쿼리 수정 — 일부 파라미터를 바꾼 뒤 진행
(번호 또는 문장으로 답해도 됩니다)
  • 1 Start now → pass the JSON query to the insane-research-main skill
  • 2 Save only → write the JSON to a file (RESEARCH/queries/{topic}_{timestamp}.json)
  • 3 Revise → collect adjustments and loop back

Quality verification rules

Verify before finalizing the query:

Task

  • Title is specific (no generic "AI Research")
  • Goal is measurable/verifiable
  • type matches the research approach

Questions

  • primary is answerable (not too broad)
  • secondary supports primary (not a derailment)
  • exclusions prevent scope creep

Constraints

  • timeframe is realistic for the topic
  • geography matches topic relevance
  • source requirements are achievable

Output

  • length matches the requested depth
  • format suits the audience

Anti-patterns

Do not generate

  • Overly broad questions ("What is AI?")
  • Infinite timeframe ("all history")
  • Conflicting constraints / generic keywords ("technology", "innovation")
  • Unmeasurable goals ("understand everything about...")

Recommended to generate

  • Specific, answerable questions ("What is the current adoption rate of AI diagnostic tools in U.S. hospitals?")
  • Realistic scope boundaries (2–3 year timeframe for fast-moving fields)
  • Concrete success criteria ("identify the top 10 tools by market share")
  • Actionable search terms ("AI radiology FDA approved 2024 2025 adoption rate")
  • Clear exclusions ("exclude consumer health apps and administrative AI")

Language adaptation

All questions/options/outputs are tailored to the user's detected language.

Handling Korean input

When the user inputs in Korean (e.g., "make me a healthcare AI research query"):

  • All questions in Korean
  • Korea-related choices in the geography options
  • Korean research DBs in the source options
  • Korean citation conventions in the output

Multilingual keywords

Generate search keywords in both the user's language and English for maximum coverage:

한국어 입력: "AI 의료 진단"
생성: ["AI 의료 진단 2026", "AI medical diagnostics 2026", "의료 AI 도입 현황", "clinical AI adoption"]

insane-research integration

The generated query is insane-research-main goes directly into the skill:

  1. The query builder outputs structured JSON
  2. The user confirms or adjusts
  3. When "start now" is selected, pass the JSON to insane-research-maindelivered to
  4. Since the requirements are already defined, Phase 1 (Question Scoping) is skipped
  5. Research starts from Phase 2 (Retrieval Planning)

Query storage location: RESEARCH/queries/{topic}_{timestamp}.json

Guardrails

  • Keep questions compact in §A numbered blocks, and do not ask what can be inferred.
  • Do not leave blanks; fill in concrete defaults.
  • Treat the query schema as the contract for the structured output.
  • Ask via chat rather than relying on widget-style prompts.

References

  • Query schema: $PLUGIN_ROOT/skills/insane-research-query/references/query_schema.json
  • Example queries: $PLUGIN_ROOT/skills/insane-research-main/examples/