菜单
Deep web research with parallel search, source validation, and evidence-based findings
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Multi-expert AI/ML specification review with scoring gate — model architecture, evaluation rigor, safety/alignment, and production readiness for AI systems and LLM apps
Multi-expert architecture review — boundaries, integration patterns, failure modes, evolvability
Multi-expert business analysis with advisory recommendations (no scoring gate)
Multi-expert intelligence pipeline review — discovery quality, ingestion resilience, scoring validity, platform compliance, taxonomy coherence, cost efficiency
Multi-expert mobile/native app specification review with scoring gate — Android, iOS, Swift, SwiftUI
Multi-expert personal-development review with scoring gate — learnability, adoption, human-centeredness, and capability impact for talent/learning/AI-augmentation designs
| name | sh-research |
| description | Deep web research with parallel search, source validation, and evidence-based findings |
Multi-query web research with source validation and structured evidence synthesis. Produces a research report -- no implementation.
Execute multiple searches concurrently using WebSearch:
Query 1: [primary question, direct terms]
Query 2: [alternative phrasing or related angle]
Query 3: [specific sub-question]
For each result, use WebFetch to extract key content from the most relevant pages.
Compile findings into structured report:
## Research: [Topic]
### Summary
[2-3 sentence executive summary]
### Findings
#### [Sub-question 1]
- Finding with [source citation]
- Confidence: High/Medium/Low
#### [Sub-question 2]
...
### Contradictions / Open Questions
- [Areas where sources disagree or data is insufficient]
### Sources
1. [Title] - [URL] - [Date] - [Relevance note]
| Level | Queries | Hops | Output |
|---|---|---|---|
| quick | 2-3 | 1 | Summary paragraph |
| standard | 4-6 | 2-3 | Structured report |
| deep | 8-12 | 3-5 | Detailed analysis with full citations |
Will: Search the web, extract content, cross-validate, produce cited report. Will not: Implement findings, make architectural decisions, access paywalled content, present uncited claims as fact.
After research, user decides action. Use findings to inform implementation or design decisions.