AgentSkillsCN

screening

全面筛查制裁与PEP名单,覆盖12+个数据库(OFAC、欧盟、英国、联合国、国际刑警组织),同时进行误报评估、人工审核决策关口,并确保符合监管要求的合规性文档。

SKILL.md
--- frontmatter
name: screening
description: >
  Comprehensive sanctions and PEP screening against 12+ databases (OFAC, EU, UK, UN, INTERPOL)
  with false positive assessment, HITL decision gates, and regulatory-compliant documentation.
user-invocable: true
category: compliance
estimated-duration: 15 minutes

Screening Skill - v1.0.0

Comprehensive sanctions and PEP screening with multi-database search, false positive assessment, and HITL decision gates.

Overview

Screening is a mandatory component of all KYC workflows. This skill provides standalone screening capability for ad-hoc checks, batch processing, and integration into onboarding/refresh workflows.

Screening Databases (12+ Sources)

Sanctions Lists

  1. OFAC SDN - US Specially Designated Nationals
  2. EU Consolidated Sanctions List - European Union
  3. UK HM Treasury - UK Financial Sanctions
  4. UN Consolidated List - United Nations Security Council
  5. Interpol - Red and Yellow Notices
  6. OpenSanctions - Open-source aggregated sanctions data

PEP Databases

  1. EveryPolitician - Global political figures
  2. Wikidata - Political and public figures
  3. UK Parliament - Register of Members' Interests
  4. CIA World Factbook - National leadership
  5. Electoral Commission databases - Candidates and office holders

Criminal/Legal

  1. UK Court records - Criminal convictions, CCJs
  2. Companies House - Insolvency records, disqualified directors

Screening Process

Step 1: Identify Screening Targets

  • Customer (individual or entity)
  • All beneficial owners (>25% threshold)
  • Directors and senior management (entities)
  • Family members and close associates (PEP cases)

Step 2: Execute Searches

For each target, search against all applicable databases:

  • Record search date, platform, and search terms
  • Document all matches (including potential false positives)
  • Screenshot or save evidence

Step 3: False Positive Assessment [HITL Checkpoint]

When a potential match is found:

FactorAssessment
Name similarityExact / Partial / Common name
Date of birthMatch / Mismatch / Unknown
NationalityMatch / Mismatch
LocationMatch / Mismatch
PhotographMatch / Mismatch / Not available
ContextSupports match / Against match

Decision:

  • TRUE MATCH: Immediate escalation to Manager + MLRO
  • FALSE POSITIVE: Document reasoning and evidence, proceed
  • UNCERTAIN: Escalate for senior review

Step 4: Generate Report

  • Screening summary with all targets and results
  • False positive assessment documentation (if applicable)
  • Escalation brief (if true match or uncertain)
  • Recommendations and next steps

Batch Screening

For processing multiple alerts:

code
Analyze [X] screening alerts from uploaded CSV.
For each alert:
- Assess match quality
- Check for false positive patterns
- Research as needed
- Provide recommendation (CLEAR/ESCALATE)
Create Excel report with all analyses.

Regulatory Basis

  • OFAC compliance (US)
  • EU Sanctions Regulations
  • UK Money Laundering Regulations 2017
  • UN Security Council Resolutions
  • AMLD5 screening requirements

Version: 1.0.0 Author: Vyayasan