Shellcheck directives allow you to control how shellcheck works, and take the form of comments in files:

hexToAscii() {
  # shellcheck disable=SC2059
  printf "\x$1"
}

Supported directives are

disable

Prevent shellcheck from processing one or more warnings:

# shellcheck disable=code[,code...]
statement_where_warning_should_be_disabled

source

Tell ShellCheck where to find a sourced file (since 0.4.0):

# shellcheck source=src/examples/config.sh
. "$(locate_config)"

shell

Specify the shell for a script (similar to the shebang, if you for any reason don't want to add one) (since 0.4.5):

# shellcheck shell=sh
echo foo &> bar

Directives that replace or are immediately after the shebang apply to the entire script. Otherwise, they are scoped to the command that follows it (including compound commands like function definitions, loops and case statements). A directive may only be applied to a complete command, and can not be used immediately preceding an else block or individual case branch:

# Directive VALID here, applies to whole `case`
case $1 in
  # Directive INVALID, `-v)` is not a complete command
  -v)
    # Directive VALID here, applies to whole `if`
    if [ "$v" ]
    then
       # Directive VALID here, applies to `die ..` command
       die "Only one -v allowed"
    # Directive INVALID here, `else` is not a complete command
    else
      v=1
    fi ;;
 esac

There is no support for scoping a directive to the first structure of the script. In these cases, use a dummy command true or : and then add directives, such as

# This directive applies to the entire script
# shellcheck disable=2086
true

# This directive only applies to this function
# shellcheck disable=2043
f() {
  ...
}

Silencing parser errors is purely cosmetic; any parser error found will still stop ShellCheck at the point of the error.