diff --git a/kstest-converter b/kstest-converter
new file mode 100755
index 00000000..867a32dc
--- /dev/null
+++ b/kstest-converter
@@ -0,0 +1,244 @@
+#!/bin/python3
+
+import argparse
+import json
+import logging
+import os
+import re
+
+SKIPS = [
+    # we haven't really figured out how to do NFS in openQA yet
+    'nfs-repo-and-addon',
+    # the prep for this one probably needs to be in fedora_openqa_schedule
+    'liveimg',
+    # i'll figure this one out later
+    'basic-ostree',
+    # FIXMEs:
+    # keyboard - changes keyboard layout, fucks with openQA's script runner
+    # hostname - changes hostname, breaks 'text_console_login' needle
+    # packages-and-groups-1 - boots to GUI login
+    # encrypt-device - need to set ENCRYPT_PASSWORD, also 'kickstart insufficient'?
+]
+
+SUBSTS = [
+    ('@KSTEST_URL@', '--mirrorlist=https://mirrors.fedoraproject.org/mirrorlist?repo=fedora-$releasever&arch=$basearch'),
+    ('@KSTEST_FTP_URL@', 'ftp://mirror.utexas.edu/pub/fedora/linux/development/rawhide/Everything/$basearch/os/'),
+    # we need to reboot not shutdown for openQA
+    ('shutdown', 'reboot')
+]
+
+def _find_tests(dir):
+    """Find the tests to run from a directory name (should be a
+    checkout of kickstart-tests).
+    """
+    # find .ks.in files
+    ksins = [ks for ks in os.listdir(dir) if ks.endswith('ks.in')]
+    # filter tests we can't do yet
+    ksins = [ks for ks in ksins if not any(ks.startswith(skip) for skip in SKIPS)]
+    # strip .ks.in
+    return [ksin.replace('.ks.in', '') for ksin in ksins]
+
+def prep_kickstarts(indir, outdir):
+    """Produce kickstarts in 'outdir' from .ks.in files in 'indir'.
+    """
+    if not os.path.isdir(outdir):
+        raise ValueError("Output directory {0} does not exist!".format(outdir))
+    tests = _find_tests(indir)
+    if not tests:
+        raise ValueError("No tests found!")
+    for test in tests:
+        with open('{0}/{1}.ks.in'.format(indir, test), 'r') as ksinfh:
+            kstext = ksinfh.read()
+        ksout = "{0}.ks".format(test)
+        for (orig, repl) in SUBSTS:
+            kstext = kstext.replace(orig, repl)
+        with open('{0}/{1}'.format(outdir, ksout), 'w') as ksoutfh:
+            ksoutfh.write(kstext)
+
+def merge_templates(indir, baseurl, tempfile, outfile):
+    """Produce openQA test suites and job templates for all tests in
+    indir, merge them with the existing openQA templates file
+    'tempfile', and write the combined templates file as 'outfile'.
+    'baseurl' is the URL to the path where the kickstart files for
+    the tests can be found.
+    """
+    tests = _find_tests(indir)
+    if not tests:
+        raise ValueError("No tests found!")
+    testsuites = [create_testsuite(test, indir, baseurl) for test in tests]
+    with open(tempfile, 'r') as tempfh:
+        templates = json.loads(tempfh.read())
+    templates = merge_testsuites(templates, testsuites)
+    with open(outfile, 'w') as outfh:
+        outfh.write(json.dumps(templates, sort_keys=True, indent=4,
+                               separators=(',', ': ')))
+
+def _get_disk_settings(test):
+    """Given text of a kickstart_tests test (.sh file) as test, return
+    a list of appropriate openQA settings dicts for hard disks.
+    """
+    # most prepare_disks just create empty disks of a given size in
+    # a standard way. in practice it's good enough to just create the
+    # right number of disks at openQA's standard size (10GB), that'll
+    # make things work. This is a rough assumption that may break in
+    # future. A missing openQA feature here is you can't tell it to
+    # create multiple clean disk images with different sizes.
+    settings = []
+    simpre = re.compile(r'qemu-img create -q -f qcow2 \$\{(tmp|disk)dir\}/disk-.\.img \d+G')
+    numdisks = len(simpre.findall(test))
+
+    # the one tricky case that exists so far is the driverdisk case.
+    # it gets created elsewhere. here we just point to it.
+    if 'mkdud.py' in test:
+        numdisks += 1
+        settings.append({'key': 'HDD_{0}'.format(str(numdisks)), 'value': "driverdisk.img"})
+
+    if numdisks > 0:
+        settings.append({'key': 'NUMDISKS', 'value': str(numdisks)})
+
+    return settings
+
+def create_testsuite(test, path, baseurl):
+    """Create an openQA 'test suite' for a given kickstart_test. test
+    is the test name, path is the directory the test files are in.
+    """
+    with open('{0}/{1}.sh'.format(path, test), 'r') as testfh:
+        sh = testfh.read()
+    with open('{0}/{1}.ks.in'.format(path, test), 'r') as ksfh:
+        ks = ksfh.read()
+    settings = _get_disk_settings(sh)
+    settings.append({'key': 'KICKSTART', 'value': '1'})
+    # just one test checks for RESULT in /home instead of /root
+    if '/home/RESULT' in sh:
+        settings.append({'key': 'POSTINSTALL', 'value': 'kstest_home'})
+    else:
+        settings.append({'key': 'POSTINSTALL', 'value': 'kstest_root'})
+    # for some goddamn reason there are two different root passwords.
+    rootpatt = re.compile(r'rootpw (.+)')
+    settings.append({'key': 'ROOT_PASSWORD', 'value': rootpatt.search(ks).group(1)})
+    settings.append({'key': 'GRUB', 'value': "inst.ks={0}/{1}.ks".format(baseurl.strip('/'), test)})
+    # we never want to do a user login for these.
+    settings.append({'key': 'USER_LOGIN', 'value': "false"})
+    return {'name': "kstest_{0}".format(test), 'settings': settings}
+
+def merge_testsuites(templates, testsuites, machine='64bit', arch='x86_64',
+                     distri='fedora', prio=50, flavor='kstests', version='*'):
+    """Merge some test suites (as produced by create_testsuite) into
+    'templates', which is expected to be an openQA templates file
+    parsed into a dict. Returns the merged dict.
+    """
+    for testsuite in testsuites:
+        templates['TestSuites'].append(testsuite)
+        jobt = {
+            'machine': {'name': machine},
+            'prio': prio,
+            'test_suite': {'name': testsuite['name']},
+            'product': {
+                'arch': arch,
+                'distri': distri,
+                'flavor': flavor,
+                'version': version,
+            },
+        }
+        templates['JobTemplates'].append(jobt)
+        foundprod = False
+        for product in templates['Products']:
+            if (
+                product['flavor'] == flavor and product['distri'] == distri and
+                product['version'] == version and product['arch'] == arch
+            ):
+                foundprod = True
+                break
+        if not foundprod:
+            # add a Product for the flavor
+            templates['Products'].append(
+                {
+                    'arch': arch,
+                    'distri': distri,
+                    'flavor': flavor,
+                    'name': "",
+                    'settings': [],
+                    'version': version
+                }
+            )
+    return templates
+
+def cmd_kickstarts(args):
+    """kickstarts subcommand function: produce kickstarts from .ks.in
+    files.
+    """
+    try:
+        prep_kickstarts(args.indir, args.outdir)
+    except ValueError as err:
+        sys.exit(err)
+
+def cmd_templates(args):
+    """templates subcommand function: produce openQA test suites and
+    job templates and merge into a templates file.
+    """
+    try:
+        merge_templates(args.indir, args.baseurl, args.tempfile, args.outfile)
+    except ValueError as err:
+        sys.exit(err)
+
+def parse_args():
+    """Parse arguments with argparse."""
+    parser = argparse.ArgumentParser(description=(
+        "Translation layer to convert anaconda kickstart-tests into "
+        "openQA tests. 'kickstarts' parses kickstart-tests .ks.in "
+        "files to kickstarts. 'templates' produces openQA test suites "
+        "from kickstart-tests and merges them into a pre-existing "
+        "openQA templats file."))
+    parser.add_argument(
+        '-l', '--loglevel', help="The level of log messages to show",
+        choices=('debug', 'info', 'warning', 'error', 'critical'),
+        default='info')
+
+    # This is a workaround for a somewhat infamous argparse bug
+    # in Python 3. See:
+    # https://stackoverflow.com/questions/23349349/argparse-with-required-subparser
+    # http://bugs.python.org/issue16308
+    subparsers = parser.add_subparsers(dest='subcommand')
+    subparsers.required = True
+
+    parser_kickstarts = subparsers.add_parser(
+        'kickstarts', description="Produce kickstarts from .ks.in "
+        "files and write them to a specified directory.")
+    parser_kickstarts.add_argument(
+        'indir', help="Input directory (containing .ks.in files)")
+    parser_kickstarts.add_argument(
+        'outdir', help="Output directory (where .ks files are written")
+    parser_kickstarts.set_defaults(func=cmd_kickstarts)
+
+    parser_templates = subparsers.add_parser(
+        'templates', description="Produce openQA 'test suites' and "
+        "'job templates' from anaconda-kickstarts tests and merge "
+        "them into an existing openQA templates file.")
+    parser_templates.add_argument(
+        'indir', help="Input directory (containing .ks.in and .sh files)")
+    parser_templates.add_argument(
+        'baseurl', help="URL to a directory containing .ks files (as "
+        "produced by 'kickstarts' subcommand)")
+    parser_templates.add_argument(
+        'tempfile', help="Path to openQA templates file (must be JSON "
+        "format, not Perl)")
+    parser_templates.add_argument('outfile', help="Path to output file")
+    parser_templates.set_defaults(func=cmd_templates)
+
+    return parser.parse_args()
+
+
+def main():
+    """Main loop."""
+    try:
+        args = parse_args()
+        loglevel = getattr(
+            logging, args.loglevel.upper(), logging.INFO)
+        logging.basicConfig(level=loglevel)
+        args.func(args)
+    except KeyboardInterrupt:
+        sys.stderr.write("Interrupted, exiting...\n")
+        sys.exit(1)
+
+if __name__ == '__main__':
+    main()
diff --git a/perl2json b/perl2json
new file mode 100755
index 00000000..eef92749
--- /dev/null
+++ b/perl2json
@@ -0,0 +1,10 @@
+#!/bin/perl
+
+use JSON;
+
+my $infile = shift @ARGV;
+my $outfile = shift @ARGV;
+my $info = do $infile;
+$info = JSON->new->relaxed->pretty->encode($info);
+open(my $outfh, '>', $outfile);
+print $outfh $info;
diff --git a/tests/kstest_home_postinstall.pm b/tests/kstest_home_postinstall.pm
new file mode 100644
index 00000000..96d0874b
--- /dev/null
+++ b/tests/kstest_home_postinstall.pm
@@ -0,0 +1,23 @@
+use base "installedtest";
+use strict;
+use testapi;
+
+sub run {
+    my $self=shift;
+    if (not( check_screen "root_console", 0)) {
+        $self->root_console(tty=>3);
+    }
+    validate_script_output 'cat /home/RESULT', sub { $_ =~ m/SUCCESS/ };
+}
+
+sub test_flags {
+    # without anything - rollback to 'lastgood' snapshot if failed
+    # 'fatal' - whole test suite is in danger if this fails
+    # 'milestone' - after this test succeeds, update 'lastgood'
+    # 'important' - if this fails, set the overall state to 'fail'
+    return { fatal => 1 };
+}
+
+1;
+
+# vim: set sw=4 et:
diff --git a/tests/kstest_root_postinstall.pm b/tests/kstest_root_postinstall.pm
new file mode 100644
index 00000000..0281c5fb
--- /dev/null
+++ b/tests/kstest_root_postinstall.pm
@@ -0,0 +1,23 @@
+use base "installedtest";
+use strict;
+use testapi;
+
+sub run {
+    my $self=shift;
+    if (not( check_screen "root_console", 0)) {
+        $self->root_console(tty=>3);
+    }
+    validate_script_output 'cat /root/RESULT', sub { $_ =~ m/SUCCESS/ };
+}
+
+sub test_flags {
+    # without anything - rollback to 'lastgood' snapshot if failed
+    # 'fatal' - whole test suite is in danger if this fails
+    # 'milestone' - after this test succeeds, update 'lastgood'
+    # 'important' - if this fails, set the overall state to 'fail'
+    return { fatal => 1 };
+}
+
+1;
+
+# vim: set sw=4 et: