osint-investigation

Resolve entities across public records into confidence-tiered evidence chains.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/MarbleSodas/Mavis --skill osint-investigation-marblesodas
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: osint-investigation
Source: https://github.com/MarbleSodas/Mavis/tree/main/optional-skills/research/osint-investigation
Command: npx skills add https://github.com/MarbleSodas/Mavis --skill osint-investigation-marblesodas

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill solves the challenge of turning scattered public records into a structured, cross-referenced OSINT investigation with explicit entity resolution and evidence trails.

Core Features & Use Cases

  • Public-record cross-referencing: Join SEC filings, federal contracts, lobbying disclosures, sanctions lists, offshore leak records, property documents, court records, web archives, and knowledge bases into a single investigation graph.
  • Entity resolution with confidence tiers: Match entities across heterogeneous sources using exact, fuzzy, and token-overlap strategies while preserving match confidence for downstream interpretation.
  • Evidence-chain outputs: Produce structured findings that trace every claim back to specific source rows, optionally augmented with timing-correlation test results.

Quick Start

Use the osint-investigation skill to resolve cross-source links for a target company across SEC EDGAR, USAspending, Senate lobbying, OFAC SDN, ICIJ offshore leaks, NYC property records, CourtListener case records, Wayback Machine captures, Wikipedia/Wikidata facts, and GDELT news by running it with the appropriate entity query and inspecting the generated evidence-chain JSON.

Frequently Asked Questions about osint-investigation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I cross-reference public records for due diligence across multiple data sources?

To perform due diligence, this Skill resolves entities across SEC EDGAR, USAspending, OFAC SDN, and ICIJ offshore leaks by joining records into a single structured investigation graph. It produces confidence-tiered matches and outputs findings JSON tracing claims to specific source rows.

What is the best way to screen an entity against sanctions lists and litigation history?

Sanctions screening and litigation history checks are performed by resolving target entities against OFAC SDN and CourtListener case records. The Skill generates structured evidence chains that link each finding to concrete source row evidence, providing confidence-tiered entity matches for downstream interpretation.

Can I use this for offshore-linked entity discovery and property ownership leads?

Offshore-linked entity discovery and property ownership leads are supported by querying ICIJ offshore leaks and NYC ACRIS property documents. The Skill cross-references these heterogeneous sources using exact, fuzzy, and token-overlap matching strategies to resolve entities while preserving match confidence.

Do I need external libraries or dependencies to run OSINT investigations with public records?

No external dependencies are required. The Skill implements deterministic, stdlib-only CSV acquisition and joins to process public records, ensuring that entity resolution and evidence-chain generation run without installing additional libraries or packages.

How does entity resolution work when investigating cross-source links from public records?

Entity resolution works by matching entities across heterogeneous public records using exact, fuzzy, and token-overlap strategies. The Skill preserves match confidence for downstream interpretation, producing structured findings that optionally include timing-correlation test results alongside concrete source row evidence.

What sources can I include when reconstructing a narrative from public records and web archives?

Narrative reconstruction with timing correlation draws from SEC EDGAR, Senate LDA, CourtListener, Wayback Machine captures, Wikipedia/Wikidata facts, and GDELT news. The Skill joins these sources into a single investigation graph, outputting findings JSON where each claim references concrete source row evidence.