Crypto tokens are digital assets, similar to the coins of a cryptocurrency, except that they do not have their own blockchain or distributed ledger. Rather, they are built on top of an existing one. Areas of application include their use as means of investment, as a local currency in a decentralized application, as well as means for building an ecosystem or a community. Depending on the purpose, it is common to categorize tokens into payment tokens, security tokens and utility tokens. The distinction is of interest since in most jurisdictions, security tokens are more heavily regulated than other tokens. In this paper, we present a heuristic approach towards automatic detection of security tokens from blockchain data. To this end, we first discuss several methods for the (semi-) automatic identification of token contracts. Then we attempt to identify the token type. For our analysis, we examine both the deployed bytecode and the calls to token contracts that we extract from transaction data of the Ethereum main chain up to block 9500000, mined on Feb 17, 2020.