LC_CTYPE=C sed '1s/^...//' < yourscript.
This is an encoding error that can't be seen in the script itself, but
cat -v will show three bytes of garbage at the start of the file:
$ cat -v file M-oM-;M-?#!/bin/bash echo "hello world"
The code is correct when this garbage does not appear.
Some editors may save a file with a Byte Order Mark to mark the file as UTF-8. Shells do not understand this and will give errors on the first line:
$ bash myscript myscript: line 1: #!/bin/sh: No such file or directory $ dash myscript myscript: 1: myscript: #!/bin/sh: not found
To fix it, remove the byte order mark. One way of doing this is
LC_CTYPE=C sed '1s/^...//' < yourscript. Verify that it's not there with
ShellCheck is a static analysis tool for shell scripts. This page is part of its documentation.