akiiwa-games/AgentCompass / skills/dot-subagent
dot-subagent
VerifiedLeaf-only runner for the local OpenAI Dot through dot2api. Use only after agent-compass has chosen Dot. Covers the runner, packets, code written and run on Dot's own computer, and how data reaches it (public re-fetch, text packets, the private GitHub repository) and how results come back. No routing and no spawning children.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.5 KB |
| agents/openai.yaml | config | 290 B |
| scripts/dot-subagent.cmd | script | 72 B |
| scripts/github_return.py | script | 8.0 KB |
| scripts/github_send.py | script | 18.1 KB |
| scripts/pack_chunks.py | script | 9.7 KB |
| scripts/parallel_fetch.py | script | 1.9 KB |
| scripts/remote_root.py | script | 1.7 KB |
| scripts/run_dot_task.py | script | 21.7 KB |
| scripts/test_executable_discovery.py | script | 2.7 KB |
| scripts/test_github_return.py | script | 6.2 KB |
| scripts/test_github_send.py | script | 11.8 KB |
| scripts/test_pack_chunks.py | script | 5.0 KB |
| scripts/test_parallel_fetch.py | script | 3.6 KB |
| scripts/test_remote_root.py | script | 1.6 KB |
| scripts/test_run_dot_task.py | script | 5.2 KB |
| scripts/test_unpack_files.py | script | 3.9 KB |
| scripts/unpack_files.py | script | 4.7 KB |
scripts/test_unpack_files.py
View on GitHubimport contextlib
import hashlib
import io
from pathlib import Path
import tempfile
import unittest
import unpack_files
def block(name, content, claimed=None):
sha = claimed or hashlib.sha256(content.encode('utf-8')).hexdigest()
return f'=== FILE: {name} sha256={sha} ===\n{content}=== END FILE ===\n'
class UnpackTests(unittest.TestCase):
def setUp(self):
self.directory = tempfile.TemporaryDirectory()
self.addCleanup(self.directory.cleanup)
self.out = Path(self.directory.name) / 'out'
def test_files_are_written_and_verified_including_nested_paths_and_unicode(self):
reply = 'EXECUTED: yes\n' + block('a.py', 'x = 1\n') + block('pkg/b_test.py', 'print("é—日本")\n') + 'NOTES: none\n'
results = unpack_files.unpack(reply, self.out)
self.assertEqual(results, [('a.py', 'ok'), ('pkg/b_test.py', 'ok')])
self.assertEqual((self.out / 'a.py').read_bytes(), b'x = 1\n')
self.assertEqual((self.out / 'pkg' / 'b_test.py').read_bytes(), 'print("é—日本")\n'.encode('utf-8'))
def test_the_runners_crlf_is_normalized_before_hashing(self):
reply = (block('a.py', 'x = 1\ny = 2\n')).replace('\n', '\r\n')
self.assertEqual(unpack_files.unpack(reply, self.out), [('a.py', 'ok')])
self.assertEqual((self.out / 'a.py').read_bytes(), b'x = 1\ny = 2\n')
def test_a_file_without_a_final_newline_is_accepted_when_that_is_what_was_hashed(self):
sha = hashlib.sha256(b'no newline').hexdigest()
reply = f'=== FILE: a.txt sha256={sha} ===\nno newline\n=== END FILE ===\n'
self.assertEqual(unpack_files.unpack(reply, self.out), [('a.txt', 'ok')])
self.assertEqual((self.out / 'a.txt').read_bytes(), b'no newline')
def test_a_wrong_hash_writes_nothing_unless_asked_to_keep_it_unverified(self):
reply = block('a.py', 'x = 1\n', claimed='0' * 64)
self.assertEqual(unpack_files.unpack(reply, self.out), [('a.py', 'mismatch')])
self.assertEqual(list(self.out.iterdir()), [])
other = Path(self.directory.name) / 'other'
self.assertEqual(unpack_files.unpack(reply, other, keep_unverified=True), [('a.py', 'mismatch')])
self.assertTrue((other / 'a.py.unverified').exists())
self.assertFalse((other / 'a.py').exists())
def test_unsafe_names_are_refused(self):
for name in ('../evil.py', '/etc/passwd', 'a/../../b.py', '.hidden', 'a//b.py', 'C:\\x.py', 'a\\b.py'):
with self.subTest(name=name), tempfile.TemporaryDirectory() as directory:
reply = block(name, 'x\n')
with self.assertRaises(ValueError):
unpack_files.unpack(reply, Path(directory) / 'o')
def test_a_non_empty_output_directory_is_refused_and_a_missing_end_line_is_an_error(self):
self.out.mkdir()
(self.out / 'keep.txt').write_text('mine')
with self.assertRaises(ValueError):
unpack_files.unpack(block('a.py', 'x\n'), self.out)
with self.assertRaises(ValueError):
unpack_files.parse('=== FILE: a.py sha256=' + '0' * 64 + ' ===\nx\n')
def test_the_exit_code_is_zero_only_when_every_file_verified(self):
base = Path(self.directory.name)
reply = base / 'reply.md'
with contextlib.redirect_stdout(io.StringIO()), contextlib.redirect_stderr(io.StringIO()):
reply.write_text(block('a.py', 'x\n'), encoding='utf-8', newline='\n')
self.assertEqual(unpack_files.main([str(reply), str(base / 'o1')]), 0)
reply.write_text(block('a.py', 'x\n', claimed='1' * 64), encoding='utf-8', newline='\n')
self.assertEqual(unpack_files.main([str(reply), str(base / 'o2')]), 1)
reply.write_text('no blocks here\n', encoding='utf-8', newline='\n')
self.assertEqual(unpack_files.main([str(reply), str(base / 'o3')]), 1)
self.assertEqual(unpack_files.main([str(base / 'missing.md'), str(base / 'o4')]), 2)
if __name__ == '__main__':
unittest.main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |