generated: '2026-09-19'
method: derived
source: openapi/chainaware-ai-enterprise-api-openapi.yml
docs:
- https://chainaware.ai/learn/api/fraud-detection-api.html
- https://chainaware.ai/learn/api/user-segmentation-api.html
- https://chainaware.ai/learn/prediction-mcp/tool-agents-trust-score-single.html
summary: >-
  A scoring API has a thin data model: there are no server-side resources with ids, only two addressable SUBJECTS
  — a Wallet (walletAddress + network) and a Contract/pool (contract or pair address + network) — and five read-only
  score documents computed about them. Identity is the chain address itself; there are no ChainAware-minted ids,
  no id prefixes and no cross-resource links other than the addresses embedded in a rug-pull result. The MCP-only
  products add two more subjects (Token, ERC-8004 Agent) and a Batch Job, described from the tool schemas and docs
  since no OpenAPI covers them.
identity:
  wallet: (network, walletAddress) — hex address or ENS name (docs accept vitalik.eth); no provider id
  contract: (network, contract or liquidity-pool address)
  erc8004_agent: (chain_id, agent_id) — the ERC-8004 Identity Registry key (MCP only)
  batch_job: job_id (uuid) + signature (hex) — both required, neither recoverable if lost (MCP only)
entities:
- name: FraudCheckResponse
  schema: openapi/chainaware-ai-enterprise-api-openapi.yml#/components/schemas/FraudCheckResponse
  returned_by: checkWalletFraud
  subject: Wallet
  fields:
  - message
  - walletAddress
  - chain
  - status
  - probabilityFraud
  - token
  - lastChecked
  - forensic_details
  - checked_times
  - createdAt
  - updatedAt
  - sanctionData
- name: WalletAuditResponse
  schema: openapi/chainaware-ai-enterprise-api-openapi.yml#/components/schemas/WalletAuditResponse
  returned_by: auditWalletBehaviour
  subject: Wallet
  fields:
  - message
  - walletAddress
  - status
  - probabilityFraud
  - token
  - chain
  - lastChecked
  - forensic_details
  - categories
  - riskProfile
  - segmentInfo
  - experience
  - intention
  - protocols
  - userDetails
  - riskCapability
  - recommendation
  - checked_times
  - createdAt
  - updatedAt
  - sanctionData
- name: RugPullResponse
  schema: openapi/chainaware-ai-enterprise-api-openapi.yml#/components/schemas/RugPullResponse
  returned_by: checkRugPull
  subject: Contract
  fields:
  - message
  - contractAddress
  - pairAddress
  - contractCreatorAddress
  - risk_score
  - risk_status
  - risk_indicators
  - liquidityEvent
  - status
  - probabilityFraud
  - chain
  - lastChecked
  - contractCreationTime
  - forensic_details
  - checked_times
  - createdAt
  - updatedAt
- name: WalletSegmentResponse
  schema: openapi/chainaware-ai-enterprise-api-openapi.yml#/components/schemas/WalletSegmentResponse
  returned_by: getWalletSegment
  subject: Wallet
  fields:
  - message
  - walletAddress
  - walletQuality
  - categories
  - riskProfile
  - segmentInfo
  - experience
  - intention
  - protocols
  - userDetails
  - recommendation
  - sanctionData
- name: CreditScoreResponse
  schema: openapi/chainaware-ai-enterprise-api-openapi.yml#/components/schemas/CreditScoreResponse
  returned_by: getCreditScore
  subject: Wallet
  fields:
  - message
  - creditData
request_bodies:
- name: FraudCheckRequestBody
  required:
  - network
  - walletAddress
  fields:
    network:
    - ETH
    - BNB
    - POLYGON
    - TON
    - BASE
    - TRON
    - HAQQ
    walletAddress: string
    calculate: boolean
- name: WalletAuditRequestBody
  required:
  - network
  - walletAddress
  fields:
    network:
    - ETH
    - BNB
    - BASE
    - HAQQ
    - SOLANA
    walletAddress: string
    calculate: boolean
- name: RugPullRequestBody
  required:
  - network
  - walletAddress
  fields:
    network:
    - ETH
    - BNB
    - BASE
    - HAQQ
    walletAddress: string
- name: WalletSegmentRequestBody
  required:
  - walletAddress
  - network
  fields:
    walletAddress: string
    network:
    - ETH
    - BNB
    - BASE
    - HAQQ
    - SOLANA
- name: CreditScoreRequestBody
  required:
  - network
  - walletAddress
  fields:
    network:
    - ETH
    walletAddress: string
mcp_only_entities:
- name: Token (Token Rank / Token Audit)
  key: (network, contract_address)
  fields:
  - rank
  - community strength
  - top holders
  - audit aggregate.risk_score 0-100
  - aggregate.verdict CLEAN|SUSPICIOUS|HIGH RISK|HONEYPOT|THEFT|UNVERIFIABLE
  - per-module status pass|fail|warn|n/a
  source: mcp/chainaware-ai-mcp-tools.json + https://chainaware.ai/learn/token-audit/index.html
- name: ERC-8004 Agent
  key: (chain_id, agent_id)
  fields:
  - owner_address
  - agent_wallet
  - wallet_verified
  - agent_uri
  - registered_at
  - metadata_json{type,name,description,image,active,supportedTrust[]}
  - registration{agent_name,agent_desc,fetch_status}
  - wallets[]{wallet_chain_id,wallet_address,source}
  - reputation_score
  - trust_score 0-1000
  - trust_tier
  - trust_flags[]
  source: https://chainaware.ai/learn/prediction-mcp/tool-agents-trust-score-single.html
- name: Batch Job
  key: job_id + signature
  fields:
  - total_items
  - chunks_enqueued
  - status pending|processing|partial|completed
  - completed/failed/pending counts
  - result address list
  source: https://chainaware.ai/learn/prediction-mcp/tool-predictive-fraud-batch.html
relationships:
- from: RugPullResponse
  to: Wallet
  type: belongs_to
  via: contractCreatorAddress
  note: The deployer wallet whose behavioural history feeds the rug-pull score; nullable.
- from: RugPullResponse
  to: Contract
  type: has_one
  via: pairAddress
  note: The DEX liquidity pair analysed alongside the token contract.
- from: FraudCheckResponse
  to: SanctionEntry
  type: has_many
  via: sanctionData[]
  note: Zero or more sanctions-list matches {category,name,description,url,isSanctioned}.
- from: FraudCheckResponse
  to: ForensicDetails
  type: has_one
  via: forensic_details
  note: 19 string flags (cybercrime, money_laundering, mixer, sanctioned, …).
- from: WalletAuditResponse
  to: FraudCheckResponse
  type: extends
  via: shared fields status/probabilityFraud/forensic_details
  note: >-
    The audit is a superset of the fraud check plus experience, intention (14 Prob_* signals), categories[], protocols[],
    riskProfile[], recommendation, segmentInfo, userDetails.
- from: ERC-8004 Agent
  to: Wallet
  type: has_many
  via: wallets[] / owner_address / agent_wallet
  note: >-
    MCP only; the agent trust score is computed from the owner and feeder wallets, which can themselves be scored
    with predictive_fraud.
- from: Batch Job
  to: Wallet
  type: has_many
  via: result address list
  note: >-
    MCP only; results are addresses, and the per-wallet documents are fetched with the single-wallet tools.
render: null
note: No subway/ diagram exists for this repo.