@ -340,14 +340,25 @@ Inbound group datagrams (radio-level `PAYLOAD_TYPE_GRP_DATA`, 0x06) are forwarde
```
Byte 0: 0x1B (packet type)
Byte 1: SNR (signed int8, scaled ×4 — divide by 4.0 to recover dB)
Bytes 2-3: Reserved
Bytes 2-3: Reserved (clients MUST ignore)
Byte 4: Channel Index (0-7)
Byte 5: Path Length (actual path when flooded, otherwise 0xFF)
Byte 5: Path Length (actual path length when flooded, otherwise 0xFF for direct)
Bytes 6-7: Data Type (uint16 little-endian)
Byte 8: Data Length
Bytes 9 .. 8+data_len: Payload
```
**Path bytes are not forwarded**: Only `path_len` is reported in the receive frame — the path itself is not copied to the host. There are no path bytes between byte 5 and the data_type field at bytes 6–7, regardless of `path_len`.
**Path Length semantics differ between send and receive**:
| Send | Flood the network | Direct route; the encoded path follows (low 6 bits = hash count, top 2 bits + 1 = hash size; on-wire byte count = `hash_count × hash_size`) |
| Receive | Packet arrived via direct route | Packet was flooded; this is the encoded `pkt->path_len` field as observed (no path bytes follow) |
In other words, the meaning of `0xFF` is inverted between the two directions, and on receive the field carries metadata only — never a routable path. `path_len` is an encoded byte (see `Packet::isValidPathLen` / `Packet::writePath` in `src/Packet.cpp`), not a raw byte count.
**Note**: The device may also emit `PACKET_MESSAGES_WAITING` (0x83) to notify the host that datagrams are queued; poll with `CMD_SYNC_NEXT_MESSAGE` (0x0A) to retrieve them.