Agent Skills › Abilityai/trinity › dd-sourcing

dd-sourcing

GitHub

用于尽职调查报告中的文档来源规范化。提供三级信源可信度分级、标准化引用格式及最佳实践，确保所有发现具备可追溯性、审计追踪和法律防御能力，并在输出中汇总来源统计与冲突处理逻辑。

docs/demos/vc-due-diligence/skills/dd-sourcing/SKILL.md Abilityai/trinity

Trigger Scenarios

撰写尽职调查报告引用外部数据或事实需要验证信息可信度生成带有审计追踪的分析结果

Install

npx skills add Abilityai/trinity --skill dd-sourcing -g -y

More Options

Non-standard path

npx skills add https://github.com/Abilityai/trinity/tree/main/docs/demos/vc-due-diligence/skills/dd-sourcing -g -y

Use without installing

npx skills use Abilityai/trinity@dd-sourcing

指定 Agent (Claude Code)

npx skills add Abilityai/trinity --skill dd-sourcing -a claude-code -g -y

安装 repo 全部 skill

npx skills add Abilityai/trinity --all -g -y

预览 repo 内 skill

npx skills add Abilityai/trinity --list

SKILL.md

Frontmatter

{
    "name": "dd-sourcing",
    "description": "Document sources properly for due diligence reports. Use when citing any external data, ensuring full traceability and auditability of findings."
}

Source Documentation Standards

All due diligence findings must be traceable to their sources. This enables audit trails, follow-up verification, and legal defensibility.

Source Quality Tiers

Tier 1: Primary Sources (Highest Credibility)

SEC filings (10-K, 10-Q, S-1)
Court documents and legal filings
Government databases (USPTO, state registrations)
Company's own audited financials
Direct customer/reference interviews

Tier 2: Secondary Sources (High Credibility)

Established data providers (Crunchbase, PitchBook, CB Insights)
Major news outlets (WSJ, Bloomberg, TechCrunch)
Industry analyst reports (Gartner, Forrester)
LinkedIn profiles (for employment verification)
Academic research and papers

Tier 3: Tertiary Sources (Use with Caution)

Social media posts
Blog articles
Glassdoor reviews
Wikipedia (only as starting point)
Forum discussions

Citation Format

Every fact must include a source citation:

{
  "source": {
    "name": "Full source name",
    "type": "SEC_FILING | NEWS | DATA_PROVIDER | INTERVIEW | GOVERNMENT | OTHER",
    "url": "https://... (if available)",
    "access_date": "2024-01-15",
    "publication_date": "2024-01-10",
    "tier": 1,
    "excerpt": "Relevant quote or data point"
  }
}

Best Practices

DO:

Capture URLs at time of access (pages change/disappear)
Note the access date for all web sources
Quote relevant passages, don't just link
Use multiple sources for critical claims
Prefer recent sources over older ones

DON'T:

Cite a source you haven't actually reviewed
Use company press releases as independent verification
Rely on a single source for material claims
Cite sources behind paywalls without noting limitation
Mix up "stated by company" vs "independently verified"

Source Tracking in Output

Include a sources section in every analysis:

{
  "sources_used": [
    {
      "id": "S1",
      "name": "Crunchbase - Acme Corp Profile",
      "url": "https://crunchbase.com/...",
      "access_date": "2024-01-15",
      "tier": 2,
      "used_for": ["funding_history", "employee_count"]
    }
  ],
  "source_statistics": {
    "tier_1": 3,
    "tier_2": 8,
    "tier_3": 2,
    "total": 13
  }
}