Installable skillApache-2.0

Background Check Summarization

Summarizes raw background check data into structured executive reports with risk categorization (GREEN/YELLOW/RED). Use when processing criminal records, court filings, employment/education verification, sanctions screening, or adverse media for pre-employment screening, executive due diligence, ...

BY CaseMarkUPDATED Oct 27, 2026
INTENDED FOR✳Claude Code>_Codex◈OpenClaw↗CursorInstall in a listed agent.

instructions.md

---
name: background-check-summary
language: en
description: Summarizes raw background check data into structured executive reports with risk categorization (GREEN/YELLOW/RED). Use when processing criminal records, court filings, employment/education verification, sanctions screening, or adverse media for pre-employment screening, executive due diligence, or investment-grade investigations.
---

# Background Check Summarization

Transforms raw background check data into a decision-ready summary with risk categorization and compliance flags. Output has three sections: executive summary, detailed findings, and methodology.

## Quick Start

1. Identify subject (name, DOB, SSN last 4, search scope)
2. Categorize each check type as Pass / Flag / Pending
3. Assign overall status: GREEN, YELLOW, or RED
4. Surface top 2–3 findings and any compliance flags
5. Append methodology and coverage limitations

## Output Structure

### Executive Summary

- **Subject**: Full name, DOB, SSN (last 4), search scope
- **Overall Status**: GREEN (clear) / YELLOW (requires review) / RED (adverse findings)
- **Category Table**: Pass/Flag/Pending per check type
- **Key Findings**: 2–3 most significant items
- **Compliance Flags**: Jurisdiction-specific requirements

### Detailed Findings

Present each category with findings or explicit "No records found":

| Category | Key Data Points |
|----------|----------------|
| Criminal Records (Fed/State/County) | Charges, disposition, date, jurisdiction, sentence |
| Sex Offender Registry | Status, jurisdiction |
| Civil Litigation | Case type, role, disposition, damages |
| Employment Verification | Title (claimed vs. verified), dates, departure reason |
| Education Verification | Degree (claimed vs. verified), institution, dates |
| Professional Licenses | Status, disciplinary actions |
| Sanctions/Watchlist | OFAC, BIS, state debarment |
| Adverse Media | Source, date, relevance assessment |

### Methodology & Limitations

State sources searched and gaps (jurisdictions not covered, sealed records, database lag).

## Risk Categorization

| Status | Criteria | Action |
|--------|----------|--------|
| GREEN | No adverse records, all verifications confirmed | Proceed |
| YELLOW | Minor discrepancies, old/minor offenses, items needing context | Human review required |
| RED | Significant adverse findings, verification failures, recent serious offenses | Enhanced review / adverse action |

### Escalation Factors

- **Criminal**: Felony; violence/fraud/theft-related; recency (<7 yrs); position relevance; pending charges
- **Employment**: Title inflation; date gaps >3 months; overlapping employment; employer not found
- **Education**: Degree or institution mismatch; unaccredited institution

## Compliance Checks

- **FCRA**: Adverse action requires pre-adverse notice, report copy, and Summary of Rights — always note this
- **Ban-the-Box / Fair Chance**: Flag jurisdictions restricting criminal history inquiries at specific hiring stages
- **EEOC Guidance**: Evaluate criminal history for nature of offense, time elapsed, and job-relatedness
- **State lookback limits**: Note applicable 7-year or other reportable-record limitations
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.