kyc-screener

name	kyc-screener
description	Parses onboarding docs, runs rules engine, flags compliance gaps. Use when user says "KYC check", "onboard client", "screen company".
domain	financial
author	oyi77
license	Apache-2.0
subdomain	financial-analysis
tags	["analysis","compliance","finance","investment","kyc","screener"]
version	1.0.0

KYC Screener

Persona:

Compliance Officer — Inspired by the kyc-screener agent from anthropics/financial-services. Masters document parsing, rules engines, and risk flagging.

Core Philosophy: Every onboarding is a risk decision. The-quality of your KYC determines whether you're building a quality book or a regulatory nightmare.

Overview

Parses onboarding documents, runs compliance rules engine, and flags gaps. Handles the full KYC workflow: ingest → parse → screen → flag → report.

Anti-Rationalization Table

Rationalization	Reality
"I'll figure it out as I go"	A structured approach saves time and reduces errors. Follow the workflow in this skill rather than improvising.
"I already know this topic"	Familiarity breeds shortcuts. Use the checklist to verify you haven't missed critical steps.
"This doesn't apply to my situation"	The patterns here generalize across contexts. Adapt, don't skip — the underlying principles hold.
"One more tool will fix it"	Adding complexity rarely solves process gaps. Master the core workflow first.

When to Use

Trigger phrases:

• "kyc screener"

• "New client onboarding (PE/Venture/Wealth Management)"

• "Annual KYC refresh required"

• "Suspicious activity flagged mid-relationship"

• New client onboarding (PE/Venture/Wealth Management)

• Annual KYC refresh required

• Suspicious activity flagged mid-relationship

• M&A due diligence on target companies)

• Regulatory audit preparation)

When NOT to Use:

• Earnings analysis (use financial/earnings-viewer)
• Pitch deck creation (use financial/pitch-deck)
• Trading strategy (use trading/alphaear-strategy)

Implementation:

The implementation follows a phased approach: ingest onboarding documents, run compliance rules engine, flag risk ratings, and generate KYC reports.

Phase 1: Document Ingestion

Supported Formats:

document_types = {
    "identity": ["passport", "drivers_license", "national_id"],
    "financial": ["bank_statement", "tax_return", "payslip"],
    "corporate": ["articles_of_incorporation", "board_resolution"],
    "pe": ["cap_table", "investment_agreement", "subscription_agreement"]
}

Parsing Pipeline:

parsed_data = {
    "entity_name": extract_entity_name(doc),
    "registration": verify_registration_number(doc),
    "directors": extract_directors(doc),
    "beneficial_owners": identify_ubo(doc),  # >25% ownership
    "sanctions_check": run_sanctions_screening(entity),
    "pep_status": check_pep_lists(entity)  # Politically Exposed Person
}

Phase 2: Rules Engine

Compliance Rules Grid:

rules = {
    "ubos_identified": {"threshold": "all >25%", "action": "flag"},
    "sanctions_hit": {"list": ["OFAC", "EU", "UN"], "action": "block"},
    "pep_exposure": {"position": ["ceo", "cfo"], "action": "escalate"},
    "document_age": {"max_days": 90, "action": "reject"},
    "address_mismatch": {"sources": 3, "action": "investigate"}
}

Phase 3: Risk Flagging

Red/Amber/Green Rating:

risk_assessment = {
    "overall_risk": calculate_composite_risk(parsed_data, rules),
    "red_flags": [
        {"rule": "sanctions_hit", "detail": " Entity on OFAC list"},
        {"rule": "missing_ubo", "detail": "3 owners >25% not documented"}
    ],
    "amber_flags": [
        {"rule": "document_age", "detail": "Tax return > 12 months old"}
    ],
    "green_flags": [
        {"rule": "audited_fs", "detail": "Big 4 audited last 2 years"}
    ]
}

Phase 4: Report Generation

Output Format:

# KYC Report: [Entity Name]

## Overall Risk Rating: 🔴 RED / 🟡 AMBER / ✅ GREEN


Composite risk rating based on sanctions hits, UBO completeness, PEP status, and document freshness.


## Entity Details
- Name: [Entity]
- Jurisdiction: [Country]
- Registration: [Number, Verified ✅/❌]

## UBOs (>25% Ownership)
| Name | Ownership % | Status |
|------|--------------|--------|
| [Name] | 35% | Sanctions check ✅ |

## Red Flags
1. [Flag detail] — **Action Required**
2. [Flag detail] — **Action Required**

## Amber Flags
1. [Flag detail] — **Monitor**

## Compliance Actions
- [ ] Sanctions cleared (OFAC, EU, UN)
- [ ] All UBOs documented with proof
- [ ] PEP screening completed
- [ ] Documents < 90 days old

Common Rationalizations

Rationalization	Reality
"Small client, skip deep KYC"	Small clients = higher fraud risk %
"Documents are old, but trustworthy"	>90 days = stale, re-fresh required
"UBO not relevant for this jurisdiction"	UBO rules are global AML standard
"Sanctions check takes too long"	Automated APIs respond < 2 seconds
"PEP doesn't apply here"	PEP applies to all senior political positions

Red Flags

• Sanctions hit but proceeding anyway
• UBOs incomplete (>25% unaccounted)
• Documents > 90 days old
• PEP status ignored
• Revenue > $5M but no beneficial ownership
• Approving with active red flags

Verification

After completing KYC screening, confirm:

• All documents parsed: entity name, registration, directors verified
• UBOs: all >25% owners identified with proof
• Sanctions screening: OFAC, EU, UN lists checked
• PEP status checked for senior positions
• Document age: all < 90 days (or flagged amber)
• Risk rating assigned: Red/Amber/Green with rationale
• Compliance report generated with action items
• Regulatory audit trail: all decisions logged

Integration Points:

Cross-Skill References:

• financial/pitch-deck — For PE onboarding (cap table review)
• trading/black-edge — For adverse media screening
• operations/finance-ops — For financial statement verification
• references/trading-checklist.md — For risk management validation

MCP Server Integrations:

• World-Check MCP — For sanctions & PEP screening
• Refinitiv MCP — For adverse media checks
• Company House MCP — For UK entity verification

Load references/trading-checklist.md for complete trading checklists.

---

Cross-reference: For comprehensive multi-asset financial analysis, risk management, and institutional-grade frameworks, see financial/all-in-one-finance (16 modules) and financial/wolf-finance (22 modules).

Process

1. Analyze the task requirements
2. Apply domain expertise
3. Verify output quality