Payload matching and packet hooks
Payload-matching keywords—content, pcre, byte_test, byte_jump, and isdataat—search the contents of traffic. In a firewall rule, these keywords are valid only at an application-layer hook, where the match binds to a specific protocol buffer such as the TLS SNI or the HTTP host. They aren't valid at a packet hook (such as tcp:all or udp:all). If you put a payload-matching keyword on a packet hook, Network Firewall rejects the rule with a message that payload matches aren't supported on a packet hook.
To match content, choose the application-layer hook where that content appears (see Common hooks for common protocols), and pair the payload keyword with the protocol buffer keyword for that field. For example, match a TLS SNI by pairing content with tls.sni at the tls:client_hello hook:
accept:flow tls:client_hello $HOME_NET any -> $EXTERNAL_NET any (tls.sni; content:"example.com"; sid:1;)
At a packet hook, use only header and flow keywords—such as dsize, ttl, tcp.flags, itype, icode, fragbits, and flow—which match packet and connection metadata rather than payload contents. To key a packet rule off a detected application protocol without matching payload, use the app-layer-protocol keyword, for example app-layer-protocol:tls. For more about packet-layer and application-layer rule tables and how payload matching binds to protocol buffers, see Firewall design