"$@" expands to all of the command line arguments as separate words. It is different from "$*", which expands to all of the arguments as a single word.

"$@" is especially useful for looping through arguments and handling arguments with spaces.

Consider we are in a script that we invoked with two arguments, like so:

$ ./script.sh "␣1␣2␣" "␣3␣␣4␣"

The variables $* or $@ will expand into $1␣$2, which in turn expand into 1␣2␣3␣4 so the loop below:

for var in $*; do # same for var in $@; do
    echo \<"$var"\>
done

will print for both

<1>
<2>
<3>
<4>

While "$*" will be expanded into "$1␣$2" which will in turn expand into "␣1␣2␣␣␣3␣␣4␣" and so the loop:

for var in "$*"; do
    echo \<"$var"\>   
done

will only invoke echo once and will print

<␣1␣2␣␣␣3␣␣4␣>

And finally "$@" will expand into "$1" "$2", which will expand into "␣1␣2␣" "␣3␣␣4␣" and so the loop

for var in "$@"; do
    echo \<"$var"\>
done

will print

<␣1␣2␣>
<␣3␣␣4␣>

thereby preserving both the internal spacing in the arguments and the arguments separation. Note that the construction for var in "$@"; do ... is so common and idiomatic that it is the default for a for loop and can be shortened to for var; do ....