diff --git a/.travis.yml b/.travis.yml
index 47df024..dc15ee3 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -21,6 +21,7 @@ script:
     - cd example && rime help
     - rime test && rime clean
     - rime htmlify_full && rime clean
+    - rime markdownify_full && rime clean
 notifications:
     email:
         on_success: never
diff --git a/example/PROJECT b/example/PROJECT
index c4d1cc7..b3ac618 100644
--- a/example/PROJECT
+++ b/example/PROJECT
@@ -3,3 +3,4 @@
 ## You can load plugins here.
 use_plugin('wikify')
 use_plugin('htmlify_full')
+use_plugin('markdownify_full')
diff --git a/rime/plugins/htmlify_full.py b/rime/plugins/htmlify_full.py
index 238d4f2..6c1588a 100644
--- a/rime/plugins/htmlify_full.py
+++ b/rime/plugins/htmlify_full.py
@@ -277,7 +277,7 @@ def _GenerateHtmlFullOne(self, problem, ui):
             cell_judge = HTMLIFY_CELL_NA
 
         # Done.
-        html = ('<tr><td>{}</td><td>{}</td><td{}</td><td{}</td>'
+        html = (u'<tr><td>{}</td><td>{}</td><td{}</td><td{}</td>'
                 '<td{}</td><td{}</td><td{}<td></tr>\n'.format(
                     title, assignees, cell_solutions, cell_input,
                     cell_output, cell_validator, cell_judge))
diff --git a/rime/plugins/markdownify_full.py b/rime/plugins/markdownify_full.py
new file mode 100644
index 0000000..a75e700
--- /dev/null
+++ b/rime/plugins/markdownify_full.py
@@ -0,0 +1,315 @@
+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+import codecs
+import getpass
+import hashlib
+import os
+import os.path
+import socket
+import sys
+
+if sys.version_info[0] == 2:
+    import commands as builtin_commands  # NOQA
+else:
+    import subprocess as builtin_commands
+
+from rime.basic import codes as basic_codes  # NOQA
+from rime.basic import consts  # NOQA
+import rime.basic.targets.project  # NOQA
+from rime.basic import test  # NOQA
+from rime.core import commands as rime_commands  # NOQA
+from rime.core import targets  # NOQA
+from rime.core import taskgraph  # NOQA
+
+MARKDOWNIFY_EMOJI_GOOD = ' :white_check_mark: '
+MARKDOWNIFY_EMOJI_NOTBAD = ' :large_blue_diamond: '
+MARKDOWNIFY_EMOJI_BAD = ' :x: '
+MARKDOWNIFY_EMOJI_NA = ' :wavy_dash: '
+
+
+def SafeUnicode(s):
+    if sys.version_info.major == 2 and not isinstance(s, unicode):  # NOQA
+        s = s.decode('utf-8')
+    return s
+
+
+def GetFileSize(dir, filename):
+    filepath = os.path.join(dir, filename)
+    if os.path.exists(filepath):
+        return '%dB' % os.path.getsize(filepath)
+    else:
+        return '-'
+
+
+def GetFileHash(dir, filename):
+    filepath = os.path.join(dir, filename)
+    if os.path.exists(filepath):
+        f = open(filepath)
+        r = f.read()
+        f.close()
+        return hashlib.md5(SafeUnicode(r).encode('utf-8')).hexdigest()
+    else:
+        return ''
+
+
+def GetMarkdownifyFileComment(dir, filename):
+    filepath = os.path.join(dir, filename)
+    if os.path.exists(filepath):
+        f = open(filepath)
+        r = f.read().strip()
+        f.close()
+        return SafeUnicode(r).replace('\n', '<br>')
+    else:
+        return ''
+
+
+class Project(targets.registry.Project):
+    @taskgraph.task_method
+    def MarkdownifyFull(self, ui):
+        markdownFull = yield self._GenerateMarkdownFull(ui)
+        codecs.open("summary.md", 'w', 'utf8').write(markdownFull)
+        yield None
+
+    @taskgraph.task_method
+    def _GenerateMarkdownFull(self, ui):
+        if not ui.options.skip_clean:
+            yield self.Clean(ui)
+
+        # Get system information.
+        rev = SafeUnicode(builtin_commands.getoutput(
+            'git show -s --oneline').replace('\n', ' ').replace('\r', ' '))
+        username = getpass.getuser()
+        hostname = socket.gethostname()
+
+        header = u'# Project Status\n\n'
+        info = u'このファイルは markdownify_full plugin により自動生成されています '
+        info += (u'(rev.%(rev)s, uploaded by %(username)s @ %(hostname)s)\n\n'
+                 % {'rev': rev, 'username': username, 'hostname': hostname})
+        footer = u''
+
+        # Generate content.
+        markdown = u'## Summary\n\n'
+        markdown += (
+            u'問題|担当|解答|入力|出力|入検|出検\n'
+            ':---|:---|:---|:---|:---|:---|:---\n'
+            )
+
+        markdownFull = u'## Detail\n\n'
+
+        results = yield taskgraph.TaskBranch([
+            self._GenerateMarkdownFullOne(problem, ui)
+            for problem in self.problems])
+        (markdownResults, markdownFullResults) = zip(*results)
+        markdown += ''.join(markdownResults) + '\n'
+        markdownFull += ''.join(markdownFullResults)
+
+        cc = os.getenv('CC', 'gcc')
+        cxx = os.getenv('CXX', 'g++')
+        java_home = os.getenv('JAVA_HOME')
+        if java_home is not None:
+            java = os.path.join(java_home, 'bin/java')
+            javac = os.path.join(java_home, 'bin/javac')
+        else:
+            java = 'java'
+            javac = 'javac'
+        environments = '## Environments\n\n'
+        environments += (
+            '- gcc\n\t- ' +
+            builtin_commands.getoutput('{0} --version'.format(cc)) +
+            '\n')
+        environments += (
+            '- g++\n\t- ' +
+            builtin_commands.getoutput('{0} --version'.format(cxx)) +
+            '\n')
+        environments += (
+            '- javac\n\t- ' +
+            builtin_commands.getoutput('{0} -version'.format(javac)) +
+            '\n')
+        environments += (
+            '- java\n\t- ' +
+            builtin_commands.getoutput('{0} -version'.format(java)) +
+            '\n')
+        environments += '\n'
+
+        errors = ''
+        if ui.errors.HasError() or ui.errors.HasWarning():
+            errors = '## Error Messages\n\n'
+            if ui.errors.HasError():
+                errors += '- ERROR:\n'
+                for e in ui.errors.errors:
+                    errors += '\t- ' + e + '\n'
+            if ui.errors.HasWarning():
+                errors += '- WARNING:\n'
+                for e in ui.errors.warnings:
+                    errors += '\t- ' + e + '\n'
+            errors += '\n'
+
+        yield (
+            header + info + markdown + environments + errors + markdownFull
+            + footer)
+
+    @taskgraph.task_method
+    def _GenerateMarkdownFullOne(self, problem, ui):
+        yield problem.Build(ui)
+
+        # Get status.
+        title = SafeUnicode(problem.title) or 'No Title'
+        assignees = problem.assignees
+        if isinstance(assignees, list):
+            assignees = ','.join(assignees)
+        assignees = SafeUnicode(assignees)
+
+        # Get various information about the problem.
+        markdownFull = '### ' + title + '\n\n'
+        solutions = sorted(problem.solutions, key=lambda x: x.name)
+        solutionnames = [solution.name for solution in solutions]
+
+        captions = [
+            name.replace('-', ' ').replace('_', ' ') for name in solutionnames]
+        markdownFull += ('|'.join(
+                         ['testcase', 'in', 'diff', 'md5'] + captions +
+                         ['Comments']) + '\n')
+        markdownFull += '|:---' * (5 + len(captions)) + '\n'
+
+        dics = {}
+        for testcase in problem.testset.ListTestCases():
+            testname = os.path.splitext(os.path.basename(testcase.infile))[0]
+            dics[testname] = {}
+        results = []
+        for solution in solutions:
+            name = solution.name
+            test_result = (yield problem.testset.TestSolution(solution, ui))[0]
+            results.append(test_result)
+            for (testcase, result) in test_result.results.items():
+                testname = os.path.splitext(
+                    os.path.basename(testcase.infile))[0]
+                dics.setdefault(testname, {})[name] = (
+                    result.verdict, result.time)
+        testnames = sorted(dics.keys())
+        lists = []
+        for testname in testnames:
+            cols = []
+            for name in solutionnames:
+                cols.append(dics[testname].get(
+                    name, (test.TestCaseResult.NA, None)))
+            lists.append((testname, cols))
+        rows = []
+        dir = problem.testset.out_dir
+        for casename, cols in lists:
+            rows.append(
+                '|'.join(
+                    [
+                        casename.replace('_', ' ').replace('-', ' '),
+                        GetFileSize(dir, casename + consts.IN_EXT),
+                        GetFileSize(dir, casename + consts.DIFF_EXT),
+                        '`%s`' % GetFileHash(dir, casename + consts.IN_EXT)[:7]
+                    ] +
+                    [self._GetMarkdownifyMessage(*t) for t in cols] +
+                    [GetMarkdownifyFileComment(dir, casename + '.comment')]
+                ) +
+                '\n')
+        markdownFull += ''.join(rows) + '\n'
+
+        # Fetch test results.
+        # results = yield problem.Test(ui)
+
+        # Get various information about the problem.
+        num_solutions = len(results)
+        num_tests = len(problem.testset.ListTestCases())
+        correct_solution_results = [result for result in results
+                                    if result.solution.IsCorrect()]
+        num_corrects = len(correct_solution_results)
+        num_incorrects = num_solutions - num_corrects
+        num_agreed = len([result for result in correct_solution_results
+                          if result.expected])
+        need_custom_judge = problem.need_custom_judge
+
+        # Solutions:
+        if num_corrects >= 2:
+            cell_solutions = MARKDOWNIFY_EMOJI_GOOD
+        elif num_corrects >= 1:
+            cell_solutions = MARKDOWNIFY_EMOJI_NOTBAD
+        else:
+            cell_solutions = MARKDOWNIFY_EMOJI_BAD
+        cell_solutions += '%d+%d' % (num_corrects, num_incorrects)
+
+        # Input:
+        if num_tests >= 20:
+            cell_input = MARKDOWNIFY_EMOJI_GOOD + str(num_tests)
+        else:
+            cell_input = MARKDOWNIFY_EMOJI_BAD + str(num_tests)
+
+        # Output:
+        if num_corrects >= 2 and num_agreed == num_corrects:
+            cell_output = MARKDOWNIFY_EMOJI_GOOD
+        elif num_agreed >= 2:
+            cell_output = MARKDOWNIFY_EMOJI_NOTBAD
+        else:
+            cell_output = MARKDOWNIFY_EMOJI_BAD
+        cell_output += '%d/%d' % (num_agreed, num_corrects)
+
+        # Validator:
+        if problem.testset.validators:
+            cell_validator = MARKDOWNIFY_EMOJI_GOOD
+        else:
+            cell_validator = MARKDOWNIFY_EMOJI_BAD
+
+        # Judge:
+        if need_custom_judge:
+            custom_judges = [
+                judge for judge in problem.testset.judges
+                if judge.__class__ != basic_codes.InternalDiffCode]
+            if custom_judges:
+                cell_judge = MARKDOWNIFY_EMOJI_GOOD
+            else:
+                cell_judge = MARKDOWNIFY_EMOJI_BAD
+        else:
+            cell_judge = MARKDOWNIFY_EMOJI_NA
+
+        # Done.
+        markdown = (u'{}|{}|{}|{}|{}|{}|{}\n'.format(
+                    title, assignees, cell_solutions, cell_input,
+                    cell_output, cell_validator, cell_judge))
+
+        yield (markdown, markdownFull)
+
+    def _GetMarkdownifyMessage(self, verdict, time):
+        if verdict is test.TestCaseResult.NA:
+            return MARKDOWNIFY_EMOJI_NA + str(verdict)
+        elif time is None:
+            return MARKDOWNIFY_EMOJI_BAD + str(verdict)
+        else:
+            return MARKDOWNIFY_EMOJI_GOOD + '%.2fs' % (time)
+
+
+class MarkdownifyFull(rime_commands.CommandBase):
+    def __init__(self, parent):
+        super(MarkdownifyFull, self).__init__(
+            'markdownify_full',
+            '',
+            'Markdownify full plugin',
+            '',
+            parent)
+        self.AddOptionEntry(rime_commands.OptionEntry(
+            's', 'skip_clean', 'skip_clean', bool, False, None,
+            'Skip cleaning generated files up.'
+        ))
+
+    def Run(self, obj, args, ui):
+        if args:
+            ui.console.PrintError(
+                'Extra argument passed to markdownify_full command!')
+            return None
+
+        if isinstance(obj, Project):
+            return obj.MarkdownifyFull(ui)
+
+        ui.console.PrintError(
+            'Markdownify_full is not supported for the specified target.')
+        return None
+
+
+targets.registry.Override('Project', Project)
+
+rime_commands.registry.Add(MarkdownifyFull)
diff --git a/tests/plugins_test/test_markdownify_full.py b/tests/plugins_test/test_markdownify_full.py
new file mode 100644
index 0000000..bbaa178
--- /dev/null
+++ b/tests/plugins_test/test_markdownify_full.py
@@ -0,0 +1,36 @@
+import unittest
+
+import mock
+
+from rime.plugins import markdownify_full
+from rime.util import struct
+
+
+class TestMarkdownifyProject(unittest.TestCase):
+    def test_do_clean(self):
+        ui = mock.MagicMock()
+        ui.options = struct.Struct({'skip_clean': False})
+        project = markdownify_full.Project('project', 'base_dir', None)
+        project.problems = []
+        project.Clean = mock.MagicMock()
+
+        task_graph = project._GenerateMarkdownFull(ui)
+        task_graph.Continue()
+
+        project.Clean.called_once_with(ui)
+
+    def test_skip_clean(self):
+        ui = mock.MagicMock()
+        ui.options = struct.Struct({'skip_clean': True})
+        project = markdownify_full.Project('project', 'base_dir', None)
+        project.problems = []
+        project.Clean = mock.MagicMock()
+
+        task_graph = project._GenerateMarkdownFull(ui)
+        task_graph.Continue()
+
+        project.Clean.assert_not_called()
+
+
+if __name__ == '__main__':
+    unittest.main()