run_token_audit¶

Triggers a deep multi-module audit of a token contract. Returns a complete risk report immediately if a cached result exists, otherwise queues an audit job and returns a job_id. Use get_token_audit_result to poll for completion when the audit is queued.

MCP Endpoint: https://prediction.mcp.chainaware.ai/sse

Supported Networks¶

Important: Token audit tools use lowercase network identifiers - different from predictive_fraud and predictive_rug_pull which use uppercase.

Identifier	Network
`eth`	Ethereum
`bsc`	BNB Smart Chain
`base`	Base
`arbitrum`	Arbitrum
`avalanche`	Avalanche
`optimism`	Optimism
`polygon`	Polygon

Workflow¶

run_token_audit is a get-or-create call:

run_token_audit
  → if audit_status == "complete"  → full result returned immediately (no second call needed)
  → if status == "queued"          → poll get_token_audit_result until audit_status == "complete"

Input Schema¶

Field	Type	Required	Description
`contract_address`	string	Yes	Token contract address to audit
`network`	string	Yes	Lowercase network identifier (e.g. `eth`, `bsc`)

Output Schema - Two Possible Shapes¶

Shape A - Cached result (audit_status: "complete"):

Returns the full risk report. Same schema as get_token_audit_result. Respond to the user directly - no second call needed.

Shape B - New audit queued:

{
  "contract_address": "string",
  "chain": "string",
  "job_id": "string",
  "status": "queued",
  "message": "string"
}

Store the job_id and begin polling get_token_audit_result with the same contract_address and network.

Code Examples¶

Node.js¶

const audit = await client.callTool({
  name: "run_token_audit",
  arguments: {
    contract_address: "0xContractAddressHere",
    network: "eth"
  }
});

if (audit.audit_status === "complete") {
  // Cached result - answer immediately
  console.log("Risk score:", audit.aggregate.risk_score);
  console.log("Verdict:", audit.aggregate.verdict);
} else {
  // Queued - store job_id and poll get_token_audit_result
  console.log("Audit queued, job_id:", audit.job_id);
}

Python¶

audit = await session.call_tool("run_token_audit", {
    "contract_address": "0xContractAddressHere",
    "network": "eth"
})

if audit.get("audit_status") == "complete":
    print("Risk score:", audit["aggregate"]["risk_score"])
    print("Verdict:", audit["aggregate"]["verdict"])
else:
    print("Queued, job_id:", audit["job_id"])

Example Agent Prompts¶

Audit this token contract: 0xContractAddress... on Ethereum
Is this BSC token safe? 0xContract...
Run a risk scan on this contract before I buy
Check if this address is a honeypot: 0x... on Base
Does this contract have hidden mint functions? 0x... on Arbitrum
Is liquidity locked for this token on Polygon?

Use Cases¶

Launchpads - block high-risk contracts before listing
DEXes - auto-scan new pools; surface risk score in the UI
Wallets - warn users before they approve or buy a risky token
Investors - due diligence before entering a position
DeFi aggregators - exclude high-risk tokens from yield routing

Error Codes¶

Code	Meaning
`400`	Malformed `contract_address` or `network`
`500`	Temporary backend failure - retry after a short delay

get_token_audit_result - poll for completed audit results
predictive_rug_pull - second opinion on rug pull probability
predictive_fraud - assess the deployer's behavioral fraud history