Enrichment Data Sourcing

What it does

When your agent uses it

• Picking a provider stack for email, phone or company enrichment
• Building or tuning a waterfall sequence to raise match rates
• Auditing credit consumption or provider performance
• Designing enrichment logic for a RevOps or data engineering team

Example prompts

• “We only have domains for these leads. Which providers should we try for emails, and in what order?”
• “Our enrichment credits ran out early this month. Audit the waterfall and suggest cheaper routing.”
• “Estimate the credit cost of enriching our whole account list with firmographics and technology stack.”

Requirements

• Python to run scripts/cost_calculator.py
• Accounts with the enrichment providers you decide to use

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Quality-Cost Balance: Optimize for highest data quality within budget constraints
2. 2Smart Routing: Route requests to providers based on input type and success probability
3. 3Waterfall Logic: Use sequential provider attempts for maximum success
4. 4Caching Strategy: Leverage cached data to reduce redundant API calls
5. 5Bulk Optimization: Process similar requests together for volume discounts

What it can do on your machine

Read from SKILL.md and the folder at commit 78e0419. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Ships 1 file in scripts/ (Python), which the agent can run.

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Enrichment Data Sourcing loads about 2.4k tokens when it runs, and up to ~2.5k if it reads all its reference files. Until then it costs about 33 tokens; SKILL.md has 629 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~33

When it runs · the whole SKILL.md, loaded when a task matches

~2.4k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~2.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from gtmagents/gtm-agents at commit 78e0419, republished under its Apache-2.0 licence (© gtmagents). 629 words, ~2,420 tokens.

name

data-sourcing

description

Optimize provider selection, routing, and credit usage across 150+ enrichment sources for company/contact intelligence.

Tier 0 (Free): Native operations, cached data, manual inputs
Tier 1 (0.5 credits): Validation, verification, basic lookups
Tier 2 (1-2 credits): Standard enrichments (Apollo, Hunter, Clearbit)
Tier 3 (2-3 credits): Premium data (ZoomInfo, technographics, intent)
Tier 4 (3-5 credits): Enterprise intelligence (PitchBook, custom AI)
Tier 5 (5-10 credits): Specialized services (video generation, deep AI research)

# Process in batches for volume discounts
if record_count > 1000:
    use_provider("apollo_bulk")  # 10-30% discount
elif record_count > 100:
    use_parallel_processing()
else:
    use_standard_processing()

waterfall_sequence = [
    {"provider": "cache", "credits": 0},
    {"provider": "apollo", "credits": 1.5, "stop_if_success": True},
    {"provider": "hunter", "credits": 1.2, "stop_if_success": True},
    {"provider": "bettercontact", "credits": 3, "stop_if_success": True},
    {"provider": "ai_research", "credits": 5, "last_resort": True}
]

Priority: Success rate over cost
Sequence:
  1. BetterContact (aggregates 10+ sources)
  2. ZoomInfo (if enterprise)
  3. Apollo + Hunter + RocketReach
  4. AI web research
Expected Success: 95%+
Average Cost: 8-12 credits

Priority: Good success with reasonable cost
Sequence:
  1. Apollo.io
  2. Hunter (if domain match)
  3. RocketReach (if name match)
  4. Stop or continue based on confidence
Expected Success: 80%
Average Cost: 3-5 credits

Priority: Minimize cost
Sequence:
  1. Cache check
  2. Hunter (domain only)
  3. Free sources (Google, LinkedIn public)
  4. Stop at first result
Expected Success: 60%
Average Cost: 1-2 credits

def calculate_data_quality_score(data, sources):
    score = 0
    
    # Multi-source validation (30 points)
    if len(sources) > 1:
        score += min(len(sources) * 10, 30)
    
    # Data completeness (30 points)
    required_fields = ["email", "phone", "title", "company"]
    score += sum(10 for field in required_fields if data.get(field))
    
    # Verification status (20 points)
    if data.get("email_verified"):
        score += 10
    if data.get("phone_verified"):
        score += 10
    
    # Recency (20 points)
    days_old = get_data_age(data)
    if days_old < 30:
        score += 20
    elif days_old < 90:
        score += 10
    
    return score

# Combine AI and traditional providers
def hybrid_enrichment(company):
    # Fast, cheap base data
    base = clearbit_lookup(company)
    
    # AI for missing pieces
    if not base.get("description"):
        base["description"] = ai_generate_description(company)
    
    # Premium for high-value
    if is_enterprise_account(base):
        base.update(zoominfo_enrich(company))
    
    return base

# Enrich in stages based on engagement
def progressive_enrichment(lead):
    # Stage 1: Basic (on import)
    if lead.stage == "new":
        return basic_enrichment(lead)  # 1-2 credits
    
    # Stage 2: Engaged (opened email)
    elif lead.stage == "engaged":
        return standard_enrichment(lead)  # 3-5 credits
    
    # Stage 3: Qualified (booked meeting)
    elif lead.stage == "qualified":
        return comprehensive_enrichment(lead)  # 10+ credits

// JavaScript/Node.js template
const enrichContact = async (name, company) => {
  // Check cache first
  const cached = await checkCache(name, company);
  if (cached) return cached;
  
  // Try providers in sequence
  const providers = ['apollo', 'hunter', 'rocketreach'];
  
  for (const provider of providers) {
    try {
      const result = await callProvider(provider, {name, company});
      if (result.email) {
        await saveToCache(result);
        return result;
      }
    } catch (error) {
      console.log(`${provider} failed, trying next...`);
    }
  }
  
  // Fallback to AI research
  return await aiResearch(name, company);
};

© gtmagents, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (scripts, references) in plugins/data-enrichment-master/skills/data-sourcing of gtmagents/gtm-agents.

• SKILL.md
• references/provider_cheat_sheet.md
• scripts/cost_calculator.py

Open the folder on GitHubat commit 78e0419

Used in 2 other repositories

We found 3 copies of this SKILL.md (exact, near-identical or edited) in other folders, from 2 other GitHub owners. This page covers the copy in gtmagents/gtm-agents, which our catalogue first saw on October 7, 2026.

• Microck/ordinary-claude-skillsskills_all/data-sourcingskills_categorized/payment/data-sourcing
• criptogus/agent-evolve-network.claude/skills/data-sourcing