Which integrations are available for this server?

Provides comprehensive token safety analysis for ERC-20 tokens on Ethereum mainnet, including honeypot detection, hidden mint identification, blacklist checks, liquidity analysis, and ownership verification through both on-chain oracle queries and off-chain scanning.

Supports token safety scoring and risk flag detection on Optimism, featuring a live deployed oracle contract for gasless on-chain safety queries (getSafetyScore/isSafe) and analysis of tokens for 17 scam patterns including honeypots and unverified source code.

Enables token safety verification and DeFi risk assessment on Polygon, allowing AI agents to check ERC-20 contracts for vulnerabilities such as fee manipulation, trading pauses, proxy risks, and LP lock status across the network.

ERC Token Safety Score

The open standard for token safety scoring on EVM chains.

Any smart contract, wallet, DEX, or AI agent can query whether a token is safe — using a single, standardized interface.

The Problem

Billions lost annually to scam tokens (honeypots, rug pulls, hidden mints)
No standard exists for token safety — every tool uses proprietary scoring
AI agents trading autonomously have no way to check safety before buying
Smart contracts cannot gate interactions based on safety data

The Standard

ERC Token Safety Score defines:

Score 0-100 — universal safety rating for any ERC-20 token
16 risk flags — bitmask identifying specific risks (honeypot, hidden mint, blacklist, etc.)
On-chain interface — getSafetyScore(address token) and isSafe(address token, uint8 minScore)
Off-chain metadata — JSON schema for detailed safety reports
AI agent discovery — .well-known/token-safety-oracle.json endpoint

interface IERC_TokenSafetyScore {
    function getSafetyScore(address token) 
        external view returns (uint8 score, uint256 flags, uint256 updatedAt);
    
    function isSafe(address token, uint8 minScore) 
        external view returns (bool safe);
}

Live Oracle

Deployed and operational on Optimism mainnet:


Contract	`0x3B8A6D696f2104A9aC617bB91e6811f489498047`
Chain	Optimism (Chain ID: 10)
Tokens scored	9 major tokens (WETH, OP, USDC, USDT, DAI, VELO, wstETH, rETH, SNX, LINK)
Gas cost per query	~0 (view function)

SafeAgent MCP Server

AI agents can query token safety via the Model Context Protocol:

MCP SSE: https://cryptogenesis.duckdns.org/mcp/sse
REST API: https://cryptogenesis.duckdns.org/token/scan?address=0x...&chain=base

Tools

Tool	Description
`check_token_safety`	Full safety analysis — honeypot detection, 17 scam patterns, LP lock check. 6 EVM chains.
`get_defi_yields`	Top DeFi yields with quality grades (A-F), 17K+ pools across 30+ chains
`get_market_overview`	DeFi market analysis with TVL breakdown and risk alerts

Supported Chains

Base, Ethereum, Arbitrum, Optimism, Polygon, BSC

Detection Capabilities

Check	Method
Honeypot	Simulates real buy+sell on DEX routers (UniV2, V3, Aerodrome)
Hidden mint	Source code pattern matching
Blacklist	Source code pattern matching
Fee manipulation	Source code pattern matching
Trading pause	Source code pattern matching
Self-destruct	Source code pattern matching
Proxy/upgradeable	Source code + on-chain detection
LP lock	Checks Unicrypt, TeamFinance, PinkSale, dead addresses
Owner status	Checks if ownership is renounced
+8 more patterns	See ERC spec for full list

Risk Flags

Each bit in the flags field represents a specific risk:

Bit	Flag	Description
0	UNVERIFIED	Source code not verified
1	HONEYPOT	Sell transactions blocked
2	HIDDEN_MINT	Owner can mint unlimited
3	BLACKLIST	Owner can blacklist addresses
4	FEE_MANIPULATION	Owner can set 100% fees
5	TRADING_PAUSE	Owner can disable trading
6	PROXY_UPGRADEABLE	Owner can change logic
7	SELF_DESTRUCT	Contract can be destroyed
8	DELEGATECALL	Arbitrary code execution
9	OWNERSHIP_NOT_RENOUNCED	Owner has privileges
10	LOW_LIQUIDITY	< $10K liquidity
11	LP_NOT_LOCKED	LP tokens not locked
12	HIGH_TAX	Tax > 10%
13-15	Reserved	Future risk categories

Integration

Smart Contract

IERC_TokenSafetyScore oracle = IERC_TokenSafetyScore(0x3B8A6D696f2104A9aC617bB91e6811f489498047);

// Check before swap
require(oracle.isSafe(tokenAddress, 60), "Token safety score too low");

API

curl "https://cryptogenesis.duckdns.org/token/scan?address=0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913&chain=base"

{
  "safety_score": 90,
  "verdict": "SAFE",
  "flags": ["OWNERSHIP_NOT_RENOUNCED"],
  "scan_time_ms": 850
}

MCP (for AI agents)

Connect to https://cryptogenesis.duckdns.org/mcp/sse and call the check_token_safety tool.

License

MIT — The standard is open. Build on it.

safeagent-token-safety