Background Check Summary
Summarizes background check data (criminal records, court filings, employment history, education verification, sanctions screening, adverse media) into cohesive executive reports. Use when processing raw background investigation data to produce decision-ready summaries with risk categorization, compliance flags, and structured findings. Supports pre-employment screening, executive due diligence, and investment-grade investigations.
Details
- License
- Apache 2.0
- Language
- English
- Version
- 1
- Updated
- Jan 27, 2026
---
name: background-check-summary
description: "Summarizes background check data (criminal records, court filings, employment history, education verification, sanctions screening, adverse media) into cohesive executive reports. Use when processing raw background investigation data to produce decision-ready summaries with risk categorization, compliance flags, and structured findings. Supports pre-employment screening, executive due diligence, and investment-grade investigations."
---
# Background Check Summarization
Transform raw background check data into structured, decision-ready summaries with risk categorization and compliance considerations.
## Output Structure
Every summary includes these sections in order:
### 1. Executive Summary (Always First)
Single-page overview:
- **Subject**: Full name, DOB, SSN (last 4), search scope
- **Overall Status**: GREEN (clear) / YELLOW (requires review) / RED (adverse findings)
- **Category Table**: Pass/Flag/Pending for each check type
- **Key Findings**: 2-3 most significant items
- **Compliance Flags**: Jurisdiction-specific requirements
### 2. Detailed Findings by Category
Present each with findings or explicit "No records found":
- Criminal Records (Federal/State/County)
- Sex Offender Registry
- Civil Litigation
- Employment Verification
- Education Verification
- Professional Licenses
- Sanctions/Watchlist Screening
- Adverse Media
### 3. Methodology & Limitations
Sources searched and any limitations.
## Risk Categorization (Traffic Light System)
| 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 |
### Criminal Risk Escalation Factors
- Felony classification
- Violence, fraud, or theft-related
- Recency (< 7 years = higher weight)
- Relevance to position
- Pending charges
### Employment Verification Flags
- Title inflation ("Manager" claimed vs. "Associate" verified)
- Date gaps > 3 months or overlapping employment
- Employer not found or closed
- Concerning departure circumstances
## Compliance Requirements
### FCRA Mandates - Never Include:
- Bankruptcies > 10 years old
- Civil suits/judgments > 7 years old
- Arrests without disposition (check jurisdiction)
- Sealed/expunged records
- Unrelated medical information
**Exception**: No time limits for positions ≥ $75,000/year in most states.
### EEOC Individualized Assessment (Required for Criminal Records)
Always note these factors:
1. **Nature and gravity** of offense
2. **Time elapsed** since offense AND completion of sentence
3. **Relevance** to position sought
### Ban-the-Box States (Private Employers)
| State | Key Requirement |
|---|---|
| California | Conditional offer first; 7-year lookback; no marijuana convictions > 2 years |
| Colorado | No inquiry until interview/conditional offer |
| Connecticut | No inquiry on initial application |
| Hawaii | Conviction must relate to job duties |
| Illinois | No inquiry until interview; individualized assessment required |
| Massachusetts | No application inquiry; sealed/expunged excluded |
| New Jersey | No inquiry until first interview |
| New York | State: No application inquiry |
| NYC | Most restrictive: complete non-criminal checks first |
| Oregon | No application inquiry |
| Washington | No application inquiry; cannot advertise "no felons" |
### Healthcare-Specific Screening (Required)
Check before hire + monthly:
- OIG LEIE (Medicare/Medicaid exclusion)
- SAM.gov Exclusions
- State Medicaid Exclusion Lists (42 states maintain separate lists)
### Adverse Action Process (FCRA-Mandated)
1. **Pre-Adverse Action Notice**: Copy of report + "Summary of Rights Under FCRA"
2. **Waiting Period**: Minimum 5 business days
3. **Final Adverse Action Notice**: CRA info, dispute rights, free report within 60 days
**Enhanced by Jurisdiction**: NYC requires written Article 23-A analysis + 3 days; California requires individualized assessment letter + 5 days.
## Formatting Standards
### Criminal Records
```
Offense: [Exact charge and statute]
Jurisdiction: [County, State / Federal District]
Case Number: [Number]
Offense Date: [Date] | Filing Date: [Date]
Disposition: [Convicted/Dismissed/Deferred/etc.] on [Date]
Classification: [Felony/Misdemeanor] - [Class]
Sentence: [Details if convicted]
Time Since Offense: [X years, Y months]
Time Since Completion: [X years, Y months]
```
### Employment Verification
| Field | Reported | Verified | Match |
|---|---|---|---|
| Employer | [Claimed] | [Confirmed] | ✓/✗/— |
| Title | [Claimed] | [Confirmed] | ✓/✗/— |
| Start Date | [Claimed] | [Confirmed] | ✓/✗/— |
| End Date | [Claimed] | [Confirmed] | ✓/✗/— |
| Salary | [Claimed] | [Confirmed/N/A] | ✓/✗/— |
| Rehire Eligible | — | [Yes/No/Policy] | — |
**Verification Method**: [Phone/Written/Database]
**Notes**: [Discrepancies or context]
### Sanctions Screening
| Database | Result | Date |
|---|---|---|
| OFAC SDN List | Clear/[Match] | [Date] |
| OIG LEIE | Clear/[Match] | [Date] |
| SAM.gov Exclusions | Clear/[Match] | [Date] |
| State Medicaid | Clear/[Match] | [Date] |
| Global PEP | Clear/[Match] | [Date] |
## Handling Ambiguous Data
### Name Matching
- Note all potential matches found
- Indicate confidence: High/Medium/Low (based on DOB, SSN, address matched)
- Flag for human verification if confidence < High
### Incomplete Verifications
- Mark as "Unable to Verify" not "Failed"
- Note verification attempts made
- Recommend follow-up actions
### Conflicting Information
- Present both versions with sources
- Do not resolve—flag for human review
## API Response Format
```json
{
"subject": { "name": "", "dob": "", "ssn_last4": "" },
"report_date": "",
"overall_status": "GREEN|YELLOW|RED",
"categories": {
"criminal_federal": { "status": "CLEAR|RECORDS_FOUND|PENDING", "findings": [] },
"criminal_state": { "status": "", "findings": [] },
"employment": { "status": "VERIFIED|PARTIAL|DISCREPANCY", "verifications": [], "gaps": [] },
"education": { "status": "", "verifications": [] },
"sanctions": { "status": "CLEAR|MATCH_FOUND", "databases_checked": [], "pep_status": {} }
},
"compliance": {
"jurisdiction_flags": [{ "jurisdiction": "", "requirement": "" }],
"eeoc_assessment_required": true,
"adverse_action_triggered": false
},
"methodology": { "sources": [], "limitations": [] }
}
```
## Critical Rules
1. **Never make hiring recommendations** — provide information for human decision-makers
2. **Always include time elapsed** for criminal records (offense date AND sentence completion)
3. **Distinguish arrests from convictions** — arrests alone cannot be basis for adverse action
4. **Flag jurisdiction-specific requirements** — ban-the-box, Fair Chance Acts, lookback limits
5. **Note verification method** — database vs. court runner vs. direct contact
6. **Preserve source attribution** — every finding must trace to a specific source
7. **Explicit "no records found"** — never leave a category unaddressed