Overview
The Financial Forensics Specialist operates as an L3 strategic command agent within the Financial domain of the Prismatic Platform. This agent performs deep transaction-level forensic analysis, producing evidence-grade documentation suitable for regulatory filings, legal proceedings, and internal compliance investigations. Where the financial-crimes-detection-commander identifies potential criminal activity through pattern detection, the Financial Forensics Specialist conducts the detailed evidentiary analysis required to substantiate or refute those initial assessments.
Within the platform's 434-agent autonomous ecosystem built on the AIAD standard, this agent serves as the forensic backbone of the Financial domain. Its work product -- detailed transaction reconstructions, fund flow analyses, and evidence chains -- forms the evidentiary foundation for regulatory submissions and supports the NO MERCY, NO DOUBTS doctrine's requirement for evidence-based decision-making without compromise.
Transaction Forensics Methodology
The agent's core methodology applies forensic accounting principles to digital financial data, reconstructing transaction sequences with full provenance chains. Every financial movement is documented with source identification, timestamp verification, counterparty resolution, and contextual annotation.
Transaction reconstruction begins with the identification of primary transaction records from source systems. These records undergo integrity verification to ensure they have not been altered since origination. Verified records are then enriched with contextual data including exchange rates at the time of transaction, counterparty identity information, and regulatory classification codes.
The reconstruction process follows the "follow the money" principle, tracing financial flows from origin to destination through all intermediate steps. For complex layering schemes involving multiple jurisdictions, financial instruments, and corporate structures, the agent maintains parallel reconstruction threads that are eventually merged into a unified flow diagram. Circular flows, split-and-merge patterns, and timing anomalies are automatically flagged as indicators requiring investigative attention.
Each reconstructed transaction carries a metadata envelope documenting the source records, enrichment steps, analytical inferences, and confidence levels. This metadata envelope ensures that every element of the reconstruction can be independently verified and that the chain of evidence remains intact for legal proceedings.
Evidence-Grade Documentation
The distinction between intelligence-grade and evidence-grade documentation is central to the Financial Forensics Specialist's mission. Intelligence-grade products support operational decision-making and may include probabilistic assessments, pattern hypotheses, and tentative conclusions. Evidence-grade products must meet stricter standards suitable for regulatory or judicial scrutiny.
Evidence-grade documentation requirements include complete source attribution for every factual claim, clear separation between observed facts and analytical inferences, explicit documentation of methodology and analytical assumptions, identification of limitations and alternative interpretations, and chronological integrity of the evidence chain.
The agent enforces these requirements through structured templates that prevent analysts from omitting required documentation elements. Template sections include executive summary, scope and methodology, factual findings (with source citations), analytical conclusions (with supporting reasoning), limitations and caveats, and appendices with raw data references.
Fund Flow Analysis
Fund flow analysis represents one of the agent's primary analytical capabilities. This technique maps the movement of financial value through networks of accounts, entities, and instruments, revealing the structure and purpose of complex financial arrangements.
The agent constructs fund flow diagrams that visualize financial movements as directed graphs, with nodes representing accounts or entities and edges representing transfers. Graph analysis techniques identify structural patterns including fan-out distributions (one source distributing to many destinations), fan-in aggregations (many sources converging to one destination), pass-through patterns (rapid movement through intermediary accounts), and round-trip flows (funds returning to their origin through circular paths).
Each pattern type carries different investigative significance. Fan-out patterns may indicate structuring to avoid reporting thresholds. Fan-in patterns may indicate aggregation of illicit proceeds. Pass-through patterns may indicate layering through shell companies. Round-trip flows may indicate loan-back schemes or trade-based laundering.
The agent quantifies fund flow metrics including total value transferred, average holding periods at each node, velocity of movement through the network, and geographic distribution of flows. These metrics provide objective measures that support comparative analysis across cases and contribute to the platform's institutional knowledge base.
Corporate Structure Analysis
Financial forensic investigations frequently require analysis of corporate ownership structures to identify beneficial owners, control relationships, and potential shell company arrangements. The agent integrates with Czech company registries and international corporate databases to reconstruct ownership chains.
Ownership chain analysis traces control from operating entities up through holding companies, trusts, and nominee arrangements to ultimate beneficial owners. The agent handles common obfuscation techniques including multi-layered holding structures, nominee shareholders and directors, cross-ownership arrangements, and offshore jurisdictions with limited disclosure requirements.
Corporate structure analysis leverages the platform's KuzuDB graph database for relationship storage and traversal, enabling efficient navigation of complex corporate networks. The entity resolution engine resolves individual references across multiple registries, handling name variations, address changes, and role transitions.
Regulatory Framework Alignment
The agent's forensic outputs align with regulatory reporting requirements across applicable jurisdictions.
| Framework | Output Alignment | Documentation Standard |
|---|---|---|
| Czech AML (253/2008 Sb.) | Suspicious Transaction Reports for FAU | Prescribed form fields with supporting analysis |
| EU 4th/5th AML Directives | Enhanced due diligence documentation | Risk factor assessment with evidence |
| FATF Standards | STR content requirements | 40 recommendation compliance checks |
| EU Sanctions | Sanctions match documentation | Screening methodology and match analysis |
| Czech Criminal Code | Expert witness documentation | Court-admissible evidence standards |
Quality Assurance Protocol
Forensic analysis quality is maintained through a multi-stage review process that applies increasing scrutiny at each stage.
The initial analysis phase produces draft findings with preliminary confidence assessments. The verification phase subjects all factual claims to independent source verification, comparing derived conclusions against original records. The consistency phase checks analytical conclusions for internal consistency and for compatibility with known facts established in related investigations. The completeness phase ensures that all relevant evidence has been considered, including potentially exculpatory evidence that may contradict the primary analytical narrative.
Quality metrics track error rates by analysis type, measuring both factual errors (incorrect data points) and analytical errors (unsupported conclusions). Historical error patterns drive targeted training and methodology refinement.
Epistemic Framework Integration
The NABLA Infinity framework's axioms directly govern the agent's analytical practices. The Contradiction Preservation axiom requires that conflicting evidence be maintained and explicitly addressed rather than silently discarded. The Provenance Mandatory axiom ensures that every analytical claim traces back to specific source records. The Signal Plurality axiom prevents conclusions based on single-source evidence.
The Trinity Gate validation applies before any forensic finding is finalized. Structural Consistency verifies that the evidence network forms a valid directed acyclic graph. Logical Consistency confirms that conclusions follow from premises through valid reasoning. Formal Necessity ensures mathematical claims are provably correct.
Integration Architecture
| Component | Relationship | Purpose |
|---|---|---|
| Prismatic OSINT | Data collection | Financial record retrieval from public sources |
| Czech Registries | Corporate data | Company records, ownership chains, financial statements |
| Risk Scoring | Analytical input | Quantitative risk assessment contributing to forensic context |
| KuzuDB | Graph storage | Corporate and financial relationship mapping |
| PostgreSQL | Structured data | Transaction records, case files, evidence metadata |
| Report Synthesis | Output generation | Formatted forensic reports with evidence appendices |
Related Agents
- financial-crimes-detection-commander (L3) - Strategic detection and coordination of financial crime investigations, providing case referrals to the forensics specialist
- financial-intelligence-commander (L3) - Broad financial intelligence synthesis providing strategic context for forensic investigations
Connect & Contribute
Created by Tomáš Korcak (korczis) | Open Source under GHL