Build a Wildfire Smoke Asset Exposure Assessor — implementation

## Build
Process atmospheric STAC Item asset links alongside simplified aerosol optical thickness records over commercial vineyards. Track cumulative smoke exposure days to assemble objective crop taint claim evidence packages.

## User
Viticulture Risk Adjuster

## Workflow
1. Ingest vineyard polygon and search STAC catalog for atmospheric and multispectral Item links during fire events.
2. Extract aerosol optical depth asset references and cloud mask metadata from matching STAC Items.
3. Evaluate ingested normalized aerosol optical thickness records to count cumulative smoke exposure days.
4. Publish structured loss investigation reports summarizing cumulative exposure and quality risk scores.

## Interface
Web map displaying smoke plume footprints, STAC scene thumbnails, and vineyard exposure gauges.

## Data
Query public Earth observation STAC API with cached Sentinel/Landsat atmospheric asset items, mapping to simplified synthetic aerosol records.
Starter examples below are synthetic. They are not downloaded records.

## Source
[STAC specification](https://raw.githubusercontent.com/radiantearth/stac-spec/master/README.md)
Parse STAC Item properties, asset hrefs, and spatial geometry from STAC specification to locate external aerosol data.

## Tools
semantic-kernel, shapely, httpx, pandas.
Semantic Kernel orchestrates AI skills linking satellite imagery assets with domain claim rubrics and external aerosol records.
Hosted reasoning model capable of generating structured claim preparation summaries based on matched external datasets.

## 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
- Verifies STAC Item spatiotemporal intersection with vineyard boundaries.
- Calculates cumulative smoke exposure days based on associated synthetic aerosol records indicating optical depth exceeding 0.5.
- Generates standardized claim evidence dossier with direct imagery asset references and human-review flags.

## Starter Examples
```json
[
  {
    "input": {
      "vineyard_id": "VALLEY-VIN-01",
      "synthetic_aerosol_records": [
        {
          "stac_id": "AOD_20230810",
          "aod_value": 0.85,
          "asset_url": "https://data.example.org/aod_20230810.json"
        },
        {
          "stac_id": "AOD_20230811",
          "aod_value": 0.92,
          "asset_url": "https://data.example.org/aod_20230811.json"
        }
      ],
      "note": "These are simplified synthetic app records, not complete official schema files."
    },
    "expected_output": {
      "vineyard_id": "VALLEY-VIN-01",
      "smoke_days_count": 2,
      "smoke_taint_risk": "ELEVATED",
      "evidence_asset_urls": [
        "https://data.example.org/aod_20230810.json",
        "https://data.example.org/aod_20230811.json"
      ],
      "dossier_status": "READY_FOR_LAB_VERIFICATION"
    }
  },
  {
    "input": {
      "vineyard_id": "VALLEY-VIN-02",
      "synthetic_aerosol_records": [
        {
          "stac_id": "AOD_20230810",
          "aod_value": 0.15,
          "asset_url": "https://data.example.org/aod_20230810.json"
        },
        {
          "stac_id": "AOD_20230812",
          "aod_value": 0.35,
          "asset_url": "https://data.example.org/aod_20230812.json"
        }
      ],
      "note": "These are simplified synthetic app records, not complete official schema files."
    },
    "expected_output": {
      "vineyard_id": "VALLEY-VIN-02",
      "smoke_days_count": 0,
      "smoke_taint_risk": "NEGLIGIBLE",
      "evidence_asset_urls": [],
      "dossier_status": "NO_EVIDENCE_OF_EXPOSURE"
    }
  }
]
```

## Safety
Evidence dossiers compile observational records to guide laboratory micro-fermentation tests, not legal claim settlement.
Use mock external writes. Require human review before any real deployment.