Build a High-Density Proxy Director Detector and SQLite Queue — implementation

## Build
Inspect corporate officer filings and beneficial ownership statements to uncover undisclosed nominee director relationships. Package ambiguous control chains into an investigation queue with SQLite.

## User
Corporate registry compliance analyst investigating potential nominee shell arrangements.

## Workflow
1. Ingest officer appointment records and match officer identities across high-volume company filings.
2. Apply BODS v0.4 nominee arrangement heuristics to detect professional proxy directors.
3. Construct a control dataset isolating proxy links and unverified beneficial principals.
4. Queue anomalous records for analyst disposition into a SQLite database with automated discrepancy scoring.

## Interface
Queue dashboard with tabular nominee risk indicators, drill-down officer networks, and disposition buttons.

## Data
Construct 50 synthetic officer registry entries and 20 entity records with nominee control links.
Starter examples below are synthetic. They are not downloaded records.

## Source
[Beneficial ownership modelling](https://standard.openownership.org/en/latest/standard/modelling/index.html)
Uses BODS v0.4 requirements for representing nominee arrangements and controlling proxy relationships.

## Tools
dspy, pydantic, sqlite3.
Optimizes prompt chains to classify complex nominee evidence statements against structured control rules.
Instruction-tuned language model with reproducible few-shot evaluation capabilities.

## Deliverables
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

## Acceptance
- Flags officers appearing across more than 10 synthetic commercial entities without operational linkage.
- Produces a valid JSON representation of circular nominee structures.
- Assigns confidence scores to suspect nominee arrangements based on proxy declaration flags.

## Starter Examples
```json
[
  {
    "input": {
      "officer_id": "off-nominee-88",
      "officer_name": "Arthur Pendelton",
      "filings": [
        {
          "company_id": "comp-1",
          "role": "director",
          "is_nominee": true,
          "principal": null
        },
        {
          "company_id": "comp-2",
          "role": "director",
          "is_nominee": true,
          "principal": "ent-shadow"
        }
      ]
    },
    "expected_output": {
      "nominee_flag": true,
      "unresolved_principals": [
        "comp-1"
      ],
      "control_chain": [
        {
          "company": "comp-2",
          "director": "off-nominee-88",
          "principal": "ent-shadow"
        }
      ],
      "queue_action": "escalate_for_rfi"
    }
  },
  {
    "input": {
      "officer_id": "off-regular-01",
      "officer_name": "Dr. Sarah Lin",
      "filings": [
        {
          "company_id": "comp-3",
          "role": "director",
          "is_nominee": false,
          "principal": null
        }
      ]
    },
    "expected_output": {
      "nominee_flag": false,
      "unresolved_principals": [],
      "control_chain": [],
      "queue_action": "clear"
    }
  }
]
```

## Safety
Uses synthetic registry profiles. Does not publish accusations of fraud or alter corporate register states. Explicitly label normalized examples as simplified synthetic app records, not complete official schema files.
Use mock external writes. Require human review before any real deployment.