Observed arrival · 2026-09-05
VetAgent puts a brake on AI crypto trades
An MCP server checks EVM and Solana tokens for honeypots, extraction taxes, liquidity problems, and rug risk before an AI agent buys or recommends them.
Field notes
VetAgent exposes three named tools: assess_token_risk, get_token_liquidity, and find_new_hot_pools. The assessment returns a categorical verdict, a numerical score, and the individual signals that fired rather than leaving the agent to interpret a large raw-data response. Its published scorecard covers 199 sampled tokens, including an 11.3% healthy-token false-positive rate and a 21.0% unknown rate, while explicitly leaving rug-detection recall blank because the available sample contained no dead tokens.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue