First, like the others have mentioned you have no bridging loop here due to running a Portchannel. That said, running STP is still fine. Let me clear some confusions on how these commands work on Cisco switches.
spanning-tree portfast trunk
This command is supposed to be run on trunk ports towards non bridging devices, such as a server with multiple VLANs or a router. This command should not be run on trunks towards switches because the port will bypass the listening and learning phase which could potentially create a bridging loop.
If you have an interface configured like this:
interface x/x
spanning-tree portfast
spanning-tree bpdufilter enable
spanning-tree bpduguard enable
BPDU guard will never kick in because BPDU filter is filtering both the outgoing and incoming BPDUs. This also means that the port can never lose its Portfast status which it would normally do if BPDUs were received inbound. If you remove the filter then BPDU guard will kick in and shutdown the port if a BPDU is received. This is done before the port can lose its Portfast operatational state so basically the port will always operate in Porfast operational mode.
If you apply the commands globally instead:
spanning-tree portfast default
spanning-tree portfast bpdufilter default
spanning-tree portfast bpduguard default
The first command enables Portfast on all access ports.
When BPDU filter is applied globally, the difference is that it sends out 11 BPDUs before going silent. Because normally one BPDU is sent out every 2 seconds and the default MaxAge is 20 seconds that means that if there is a device at the other end that can process BPDUs, at least one BPDU would be received when the old BPDU (if there was one) has expired.
If a BPDU is received inbound when BPDU filter is applied globally then the port stops filtering and it will lose its Portfast status.
The BPDU guard default command will only apply to ports that are in a Portfast operational state.
If you combine these three commands together then what will happen is that when a BPDU is received the port loses its BPDU filter, BPDU guard can then kick in. The port will never lose its Portfast operational state because the port is shutdown before.
So you see when applied to the interface BPDU guard can never kick in but if you apply it globally it can.
If you run just Portfast globally and BPDU filter globally then if a BPDU comes in, the port loses the filter and loses the Portfast operational state and will operate as a normal port.
Use spanning-tree portfast, and spanning-tree bpduguard enable on each switchport. Root guard is unnecessary on ports with this configuration, because bpduguard will err-disable the port when you receive any bpdus.
Use root guard on links to other switches, which are not planned for the primary or secondary stp root role.
Best Answer
IEEE BPDUs are always sent untagged. If you check out the current 802.1Q-2018 (free access after registration), the BPDU form is detailed in Clause 14 and MSTP itself in Clause 13.
Even with Multiple Spanning Tree Protocol (MSTP), a single untagged BPDU holds all information about all configured MST instances (max 64 MSTIs). MCIDs are used to make sure all bridges talk about the same VLANs/trees.
BPDUs are not forwarded, so it's easier to visualize them as not belonging to any VLAN. They're simply transmitted across a link without a tag.