# KCG Entry Format for Arweave Storage

Each KCG Entry represents either an Entity or a Relation and is stored immutably on Arweave/IPFS. The entries are designed to be content-addressable and verifiable, ensuring transparency and data integrity.

# Entity Entry Example (JSON format):

{
  "type": "Entity",
  "id": "kcg:0xabc123",
  "label": "Electric Vehicle",
  "attributes": {
    "category": "Product",
    "manufacturer": "Tesla",
    "region": "Global"
  },
  "evidence": {
    "source": "DoD-Query from Claude 3",
    "confidence": 0.92,
    "created_at": "2025-05-14T12:34:56Z",
    "created_by": "DoD-Agent:Node42"
  }
}

# Relation Entry Example (JSON format):

{
  "type": "Relation",
  "id": "kcg:0xdef456",
  "source": "kcg:0xabc123",
  "target": "kcg:0xdef789",
  "relation": "depends_on",
  "context": "Global Battery Supply Chain",
  "evidence": {
    "source": "DoD-Query from GPT-4 Turbo",
    "confidence": 0.89,
    "created_at": "2025-05-14T12:36:00Z",
    "created_by": "DoD-Agent:Node42"
  }
}

# Arweave Transaction Tags:

Each entry is accompanied by Arweave tags for indexing and searchability:

Tag Key	Example Value
`App`	`KnowledgeCacheGraph`
`Type`	`Entity` / `Relation`
`Relation`	`depends_on` (for Relation type)
`SourceModel`	`GPT-4`, `Claude 3`
`Category`	`SupplyChain`, `Product`, etc.
`Confidence`	`0.89`
`CreatedBy`	`DoD-Agent:Node42`

# Integrity and Linking:

The id field is derived from the Arweave transaction ID (TXID), ensuring content-addressability.
source and target in Relation entries refer to other Entity IDs, maintaining graph structure.
Provenance data in evidence ensures transparency of origin and quality.

# Storage Optimizations:

Bundling multiple entries in a single Arweave transaction via Bundlr to optimize costs.
CID linkage for efficient DAG-based traversal.

This format provides a robust, scalable method for storing and retrieving validated knowledge in a decentralized, immutable graph structure.