Merge PR #1 from RJS

This commit is contained in:
K. Hodges 2026-05-24 00:55:02 -07:00
commit 03438e1e8a
8 changed files with 310 additions and 87 deletions

View File

@ -122,6 +122,24 @@ Semantic context stage:
This stage builds a lightweight repository index of files, Python symbols, imports, and tests, then writes compact relevant snippets for the current task. It is keyword based with symbol-aware scoring, so it works without a vector database or network dependency. This stage builds a lightweight repository index of files, Python symbols, imports, and tests, then writes compact relevant snippets for the current task. It is keyword based with symbol-aware scoring, so it works without a vector database or network dependency.
### `on_status` Stage Routing
Instead of a single `on_fail` catch-all, use `on_status` to route each review status to a different stage:
```yaml
- id: review
type: agent_review
agent: reviewer
output: review.md
on_status:
pass: summarize
retry: implement
fail: plan
escalate: human
```
`on_status` supports `pass`, `fail`, `retry`, and `escalate` keys. For `pass`, it overrides sequential progression and any agent-supplied `next_stage`. For non-pass statuses, the lookup order is: `on_status[status]``on_fail``next_stage` (agent output).
## Failure, Retry, and Resource Artifacts ## Failure, Retry, and Resource Artifacts
Failed command and validation stages write deterministic diagnostics under the task artifact directory: Failed command and validation stages write deterministic diagnostics under the task artifact directory:

View File

@ -33,6 +33,7 @@ class SafetyConfig:
allowed_commands: tuple[str, ...] allowed_commands: tuple[str, ...]
forbidden_commands: tuple[str, ...] forbidden_commands: tuple[str, ...]
allowed_env: tuple[str, ...] = () allowed_env: tuple[str, ...] = ()
skip_repo_parts: tuple[str, ...] = ()
@dataclass(frozen=True) @dataclass(frozen=True)
@ -61,7 +62,7 @@ class StageConfig:
commands: tuple[str, ...] = () commands: tuple[str, ...] = ()
output: str | None = None output: str | None = None
on_fail: str | None = None on_fail: str | None = None
on_pass: str | None = None on_status: dict[str, str] | None = None
shell: bool = True shell: bool = True
timeout_seconds: int | None = None timeout_seconds: int | None = None
working_dir: Path | None = None working_dir: Path | None = None
@ -184,6 +185,9 @@ def parse_config(raw: dict[str, Any], config_path: Path) -> NightShiftConfig:
) )
safety_raw = _require_mapping(raw["safety"], "safety") safety_raw = _require_mapping(raw["safety"], "safety")
skip_repo_parts = _string_tuple(
safety_raw.get("skip_repo_parts", []), "safety.skip_repo_parts"
)
safety = SafetyConfig( safety = SafetyConfig(
require_clean_worktree=_optional_bool( require_clean_worktree=_optional_bool(
safety_raw.get("require_clean_worktree", False), safety_raw.get("require_clean_worktree", False),
@ -195,6 +199,7 @@ def parse_config(raw: dict[str, Any], config_path: Path) -> NightShiftConfig:
safety_raw.get("forbidden_commands", []), "safety.forbidden_commands" safety_raw.get("forbidden_commands", []), "safety.forbidden_commands"
), ),
allowed_env=_string_tuple(safety_raw.get("allowed_env", []), "safety.allowed_env"), allowed_env=_string_tuple(safety_raw.get("allowed_env", []), "safety.allowed_env"),
skip_repo_parts=skip_repo_parts,
) )
agents_raw = _require_mapping(raw["agents"], "agents") agents_raw = _require_mapping(raw["agents"], "agents")
@ -393,7 +398,7 @@ def parse_config(raw: dict[str, Any], config_path: Path) -> NightShiftConfig:
commands=commands, commands=commands,
output=_optional_string(stage_raw.get("output"), f"{stage_context}.output"), output=_optional_string(stage_raw.get("output"), f"{stage_context}.output"),
on_fail=_optional_string(stage_raw.get("on_fail"), f"{stage_context}.on_fail"), on_fail=_optional_string(stage_raw.get("on_fail"), f"{stage_context}.on_fail"),
on_pass=_optional_string(stage_raw.get("on_pass"), f"{stage_context}.on_pass"), on_status=_parse_on_status(stage_raw, stage_context),
shell=_optional_bool(stage_raw.get("shell", True), f"{stage_context}.shell"), shell=_optional_bool(stage_raw.get("shell", True), f"{stage_context}.shell"),
timeout_seconds=timeout_seconds, timeout_seconds=timeout_seconds,
working_dir=Path(working_dir_raw) if working_dir_raw else None, working_dir=Path(working_dir_raw) if working_dir_raw else None,
@ -418,9 +423,12 @@ def parse_config(raw: dict[str, Any], config_path: Path) -> NightShiftConfig:
raise ConfigError( raise ConfigError(
f"Config error: stage '{stage.id}' on_fail references unknown stage '{stage.on_fail}'." f"Config error: stage '{stage.id}' on_fail references unknown stage '{stage.on_fail}'."
) )
if stage.on_pass and stage.on_pass not in stage_ids: if stage.on_status:
for status_key, target in stage.on_status.items():
if target not in stage_ids:
raise ConfigError( raise ConfigError(
f"Config error: stage '{stage.id}' on_pass references unknown stage '{stage.on_pass}'." f"Config error: stage '{stage.id}' on_status.{status_key} "
f"references unknown stage '{target}'."
) )
return NightShiftConfig( return NightShiftConfig(
@ -635,3 +643,27 @@ def _string_tuple(value: Any, context: str) -> tuple[str, ...]:
if not isinstance(value, list) or not all(isinstance(item, str) and item for item in value): if not isinstance(value, list) or not all(isinstance(item, str) and item for item in value):
raise ConfigError(f"Config error: '{context}' must be a list of non-empty strings.") raise ConfigError(f"Config error: '{context}' must be a list of non-empty strings.")
return tuple(value) return tuple(value)
VALID_STATUS_KEYS = frozenset({"pass", "fail", "retry", "escalate"})
def _parse_on_status(raw: dict[str, Any], context: str) -> dict[str, str] | None:
on_status_raw = raw.get("on_status")
if on_status_raw is None:
return None
if not isinstance(on_status_raw, dict):
raise ConfigError(f"Config error: {context}.on_status must be a mapping.")
on_status: dict[str, str] = {}
for key, value in on_status_raw.items():
if key not in VALID_STATUS_KEYS:
raise ConfigError(
f"Config error: {context}.on_status invalid key '{key}'. "
f"Valid keys: {', '.join(sorted(VALID_STATUS_KEYS))}."
)
if not isinstance(value, str) or not value:
raise ConfigError(
f"Config error: {context}.on_status.{key} must be a non-empty string."
)
on_status[key] = value
return on_status

View File

@ -106,32 +106,16 @@ def parse_file_updates(text: str) -> tuple[FileUpdate, ...]:
updates.append(FileUpdate(path=path, content=content)) updates.append(FileUpdate(path=path, content=content))
if not updates: if not updates:
raise PipelineError( raise PipelineError(
"File writer error: no file blocks found. Expected FILE: path with ---CONTENT---/---END--- or fenced blocks like ```file:path.py." "File writer error: no file blocks found. Expected fenced blocks like ```file:path.to."
) )
return tuple(updates) return tuple(updates)
def _parse_delimited_file_updates(text: str) -> list[FileUpdate]: def _parse_delimited_file_updates(text: str) -> list[FileUpdate]:
updates: list[FileUpdate] = []
header_pattern = re.compile(r"(?m)^FILE:\s*(?P<path>[^\n]+)\n---CONTENT---\n")
matches = list(header_pattern.finditer(text))
for index, match in enumerate(matches):
path = match.group("path").strip().strip("`")
content_start = match.end()
next_file_start = matches[index + 1].start() if index + 1 < len(matches) else len(text)
raw_content = text[content_start:next_file_start]
end_match = re.search(r"(?m)^---END---\s*$", raw_content)
if end_match:
raw_content = raw_content[: end_match.start()]
content = raw_content.rstrip("\r\n") + "\n"
if path:
updates.append(FileUpdate(path=path, content=content))
if updates:
return updates
pattern = re.compile( pattern = re.compile(
r"(?ms)^FILE:\s*(?P<path>[^\n]+)\n---CONTENT---\n(?P<content>.*?)\n---END---\s*$" r"(?ms)^FILE:\s*(?P<path>[^\n]+)\n---CONTENT---\n(?P<content>.*?)\n---END---\s*$"
) )
updates: list[FileUpdate] = []
for match in pattern.finditer(text): for match in pattern.finditer(text):
path = match.group("path").strip().strip("`") path = match.group("path").strip().strip("`")
content = match.group("content") content = match.group("content")

View File

@ -200,8 +200,26 @@ class PipelineRunner:
retry_notes.append(f"Context update from '{stage.id}': {result.context_update}") retry_notes.append(f"Context update from '{stage.id}': {result.context_update}")
if result.status == "pass": if result.status == "pass":
pass_target_stage = result.next_stage or stage.on_pass if stage.on_status and "pass" in stage.on_status:
if stage.type in {"agent_review", "review"} and result.next_stage: target = stage.on_status["pass"]
if target not in stage_indexes:
final_status = "failed"
final_reason = (
f"Stage '{stage.id}' on_status.pass references unknown stage '{target}'."
)
break
self.logger.event(
"stage.next",
"Jumping via on_status.pass",
run_id=self.artifacts.run_id,
task_id=task.id,
stage_id=stage.id,
next_stage=target,
)
index = stage_indexes[target]
continue
if stage.type in {"agent_review", "review"}:
if result.next_stage:
self.logger.event( self.logger.event(
"stage.next_ignored", "stage.next_ignored",
"Ignoring next_stage from passing review", "Ignoring next_stage from passing review",
@ -210,12 +228,13 @@ class PipelineRunner:
stage_id=stage.id, stage_id=stage.id,
requested_next_stage=result.next_stage, requested_next_stage=result.next_stage,
) )
pass_target_stage = stage.on_pass index += 1
if pass_target_stage: continue
if pass_target_stage not in stage_indexes: if result.next_stage:
if result.next_stage not in stage_indexes:
final_status = "failed" final_status = "failed"
final_reason = ( final_reason = (
f"Stage '{stage.id}' requested unknown next stage '{pass_target_stage}'." f"Stage '{stage.id}' requested unknown next stage '{result.next_stage}'."
) )
break break
self.logger.event( self.logger.event(
@ -224,14 +243,14 @@ class PipelineRunner:
run_id=self.artifacts.run_id, run_id=self.artifacts.run_id,
task_id=task.id, task_id=task.id,
stage_id=stage.id, stage_id=stage.id,
next_stage=pass_target_stage, next_stage=result.next_stage,
) )
index = stage_indexes[pass_target_stage] index = stage_indexes[result.next_stage]
continue continue
index += 1 index += 1
continue continue
target_stage = _failure_target_stage(stage, result) target_stage = _resolve_retry_target_stage(stage, result)
analysis_note = self._write_failure_diagnostics(stage, task, result, retry_count) analysis_note = self._write_failure_diagnostics(stage, task, result, retry_count)
if analysis_note: if analysis_note:
retry_notes.append(analysis_note) retry_notes.append(analysis_note)
@ -1840,14 +1859,10 @@ def _is_malformed_review_result(result: StageResult) -> bool:
) )
def _failure_target_stage(stage: StageConfig, result: StageResult) -> str | None: def _resolve_retry_target_stage(stage: StageConfig, result: StageResult) -> str | None:
if stage.type not in {"agent_review", "review"}: if stage.type in {"agent_review", "review"} and _is_malformed_review_result(result):
return result.next_stage or stage.on_fail
if _is_malformed_review_result(result):
return None return None
if result.next_stage and result.next_stage != stage.id: return (stage.on_status or {}).get(result.status) or stage.on_fail or result.next_stage
return result.next_stage
return stage.on_fail
def _previous_continuity_review_passed(previous_outputs: dict[str, str]) -> bool: def _previous_continuity_review_passed(previous_outputs: dict[str, str]) -> bool:

View File

@ -17,7 +17,7 @@ from .safety import resolve_inside_root, resolve_project_root, validate_scoped_p
DEFAULT_MAX_BYTES = 20_000 DEFAULT_MAX_BYTES = 20_000
DEFAULT_MAX_MATCHES = 100 DEFAULT_MAX_MATCHES = 100
DEFAULT_MAX_LOOKUP_REQUESTS = 8 DEFAULT_MAX_LOOKUP_REQUESTS = 8
SKIPPED_REPO_PARTS = {".git", ".nightshift", "__pycache__", ".venv", "venv"} DEFAULT_SKIPPED_REPO_PARTS = {".git", ".nightshift", "__pycache__", ".venv", "venv"}
@dataclass(frozen=True) @dataclass(frozen=True)
@ -45,6 +45,7 @@ class RepoTools:
self.project_root, self.project_root,
safety.scoped_paths or (".",), safety.scoped_paths or (".",),
) )
self.skipped_parts = DEFAULT_SKIPPED_REPO_PARTS | set(safety.skip_repo_parts)
def list_files(self, path: str = ".", pattern: str = "*", max_files: int = 200) -> str: def list_files(self, path: str = ".", pattern: str = "*", max_files: int = 200) -> str:
root = self._resolve_scoped(path, "list_files path") root = self._resolve_scoped(path, "list_files path")
@ -57,7 +58,7 @@ class RepoTools:
relative_files = [ relative_files = [
_relative(item, self.project_root) _relative(item, self.project_root)
for item in sorted(candidates) for item in sorted(candidates)
if fnmatch.fnmatch(item.name, pattern) and not _is_skipped_repo_path(item, self.project_root) if fnmatch.fnmatch(item.name, pattern) and not _is_skipped_repo_path(item, self.project_root, self.skipped_parts)
] ]
lines = relative_files[:max_files] lines = relative_files[:max_files]
if len(relative_files) > max_files: if len(relative_files) > max_files:
@ -66,7 +67,7 @@ class RepoTools:
def read_file(self, path: str, max_bytes: int = DEFAULT_MAX_BYTES) -> str: def read_file(self, path: str, max_bytes: int = DEFAULT_MAX_BYTES) -> str:
file_path = self._resolve_scoped(path, "read_file path") file_path = self._resolve_scoped(path, "read_file path")
if _is_skipped_repo_path(file_path, self.project_root): if _is_skipped_repo_path(file_path, self.project_root, self.skipped_parts):
return f"Path is skipped for repository lookup: {path}" return f"Path is skipped for repository lookup: {path}"
if not file_path.exists() or not file_path.is_file(): if not file_path.exists() or not file_path.is_file():
return f"File not found: {path}" return f"File not found: {path}"
@ -89,7 +90,7 @@ class RepoTools:
files = [root] if root.is_file() else [item for item in root.rglob("*") if item.is_file()] files = [root] if root.is_file() else [item for item in root.rglob("*") if item.is_file()]
matches: list[str] = [] matches: list[str] = []
for file_path in sorted(files): for file_path in sorted(files):
if _is_skipped_repo_path(file_path, self.project_root): if _is_skipped_repo_path(file_path, self.project_root, self.skipped_parts):
continue continue
try: try:
text = file_path.read_text(encoding="utf-8", errors="replace") text = file_path.read_text(encoding="utf-8", errors="replace")
@ -270,9 +271,9 @@ def _relative(path: Path, root: Path) -> str:
return path.as_posix() return path.as_posix()
def _is_skipped_repo_path(path: Path, root: Path) -> bool: def _is_skipped_repo_path(path: Path, root: Path, skipped_parts: set[str]) -> bool:
try: try:
parts = set(path.relative_to(root).parts) parts = set(path.relative_to(root).parts)
except ValueError: except ValueError:
parts = set(path.parts) parts = set(path.parts)
return bool(parts & SKIPPED_REPO_PARTS) return bool(parts & skipped_parts)

View File

@ -68,6 +68,12 @@ pipeline:
- id: review - id: review
type: agent_review type: agent_review
agent: reviewer agent: reviewer
# on_fail: implement # catch-all for any non-pass status
# on_status: # per-status routing (takes priority over on_fail)
# pass: summarize
# retry: implement
# fail: plan
# escalate: human
on_fail: implement on_fail: implement
output: review.md output: review.md
@ -333,7 +339,7 @@ If you need repository context before planning, output lookup requests exactly l
lookup_requests: lookup_requests:
- tool: read_file - tool: read_file
path: relative/path.py path: relative/path.to
- tool: grep - tool: grep
path: . path: .
pattern: search_regex pattern: search_regex
@ -350,7 +356,7 @@ REAL_MODEL_IMPLEMENTER_PROMPT = """You are the implementation agent for NightShi
Output only complete file content blocks. Output only complete file content blocks.
Use one fenced block per file with this exact opening form: Use one fenced block per file with this exact opening form:
```file:relative/path.py ```file:relative/path.to
<complete file content> <complete file content>
``` ```
Do not include explanations before or after the file blocks. Do not include explanations before or after the file blocks.

View File

@ -55,39 +55,57 @@ class ConfigTests(unittest.TestCase):
with self.assertRaisesRegex(ConfigError, "on_fail references unknown stage"): with self.assertRaisesRegex(ConfigError, "on_fail references unknown stage"):
load_config(config_path) load_config(config_path)
def test_on_pass_must_reference_existing_stage(self) -> None: def test_on_status_parses_correctly(self) -> None:
with tempfile.TemporaryDirectory() as directory: with tempfile.TemporaryDirectory() as directory:
root = Path(directory) root = Path(directory)
init_project(root) init_project(root)
config_path = root / "nightshift.yaml" config_path = root / "nightshift.yaml"
config_path.write_text( text = config_path.read_text(encoding="utf-8")
config_path.read_text(encoding="utf-8").replace( text = text.replace(
"on_fail: plan", "on_pass: missing_stage", 1 " on_fail: implement\n output: review.md",
), " output: review.md\n on_status:\n pass: summarize\n retry: implement\n fail: plan",
encoding="utf-8",
)
with self.assertRaisesRegex(ConfigError, "on_pass references unknown stage"):
load_config(config_path)
def test_on_pass_loads(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
init_project(root)
config_path = root / "nightshift.yaml"
config_path.write_text(
config_path.read_text(encoding="utf-8").replace(
" output: plan.md",
" output: plan.md\n on_pass: summarize",
1,
),
encoding="utf-8",
) )
config_path.write_text(text, encoding="utf-8")
config = load_config(config_path) config = load_config(config_path)
plan_stage = next(stage for stage in config.pipeline.stages if stage.id == "plan") review_stage = next(s for s in config.pipeline.stages if s.id == "review")
self.assertEqual(plan_stage.on_pass, "summarize") self.assertEqual(review_stage.on_status, {
"pass": "summarize",
"retry": "implement",
"fail": "plan",
})
self.assertIsNone(review_stage.on_fail)
def test_on_status_rejects_invalid_key(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
init_project(root)
config_path = root / "nightshift.yaml"
text = config_path.read_text(encoding="utf-8")
text = text.replace(
" on_fail: implement\n output: review.md",
" output: review.md\n on_status:\n wat: broken",
)
config_path.write_text(text, encoding="utf-8")
with self.assertRaisesRegex(ConfigError, "on_status invalid key"):
load_config(config_path)
def test_on_status_references_unknown_stage(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
init_project(root)
config_path = root / "nightshift.yaml"
text = config_path.read_text(encoding="utf-8")
text = text.replace(
" on_fail: implement\n output: review.md",
" output: review.md\n on_status:\n fail: missing_stage",
)
config_path.write_text(text, encoding="utf-8")
with self.assertRaisesRegex(ConfigError, "on_status.fail references unknown stage"):
load_config(config_path)
def test_validate_requires_prompt_files(self) -> None: def test_validate_requires_prompt_files(self) -> None:
with tempfile.TemporaryDirectory() as directory: with tempfile.TemporaryDirectory() as directory:
@ -371,6 +389,39 @@ class ConfigTests(unittest.TestCase):
with self.assertRaisesRegex(ConfigError, "non-command stage 'plan'"): with self.assertRaisesRegex(ConfigError, "non-command stage 'plan'"):
load_config(config_path) load_config(config_path)
def test_on_status_empty_key_fails(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
init_project(root)
config_path = root / "nightshift.yaml"
text = config_path.read_text(encoding="utf-8")
text = text.replace(
" on_fail: implement\n output: review.md",
" output: review.md\n on_status:\n pass: ",
)
config_path.write_text(text, encoding="utf-8")
with self.assertRaisesRegex(ConfigError, "must be a non-empty string"):
load_config(config_path)
def test_on_fail_fallback_when_on_status_does_not_cover_status(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
init_project(root)
config_path = root / "nightshift.yaml"
text = config_path.read_text(encoding="utf-8")
text = text.replace(
" on_fail: implement\n output: review.md",
" output: review.md\n on_status:\n pass: summarize\n on_fail: implement",
)
config_path.write_text(text, encoding="utf-8")
config = load_config(config_path)
review_stage = next(s for s in config.pipeline.stages if s.id == "review")
self.assertEqual(review_stage.on_status, {"pass": "summarize"})
self.assertEqual(review_stage.on_fail, "implement")
if __name__ == "__main__": if __name__ == "__main__":
unittest.main() unittest.main()

View File

@ -105,29 +105,145 @@ class PipelineRunnerTests(unittest.TestCase):
) )
self.assertIn("Modified Files", (root / ".nightshift" / "runs" / "test-run" / "run-summary.md").read_text(encoding="utf-8")) self.assertIn("Modified Files", (root / ".nightshift" / "runs" / "test-run" / "run-summary.md").read_text(encoding="utf-8"))
def test_on_pass_jumps_to_configured_stage(self) -> None: def test_on_status_routes_pass_to_target(self) -> None:
with tempfile.TemporaryDirectory() as directory: with tempfile.TemporaryDirectory() as directory:
root = Path(directory) root = Path(directory)
_write_common_files(root) _write_common_files(root)
stages = ( stages = (
StageConfig(id="first", type="agent", agent="planner", output="first.md", on_pass="third"), StageConfig(id="plan", type="agent", agent="planner", output="plan.md"),
StageConfig( StageConfig(
id="second", id="review",
type="command", type="agent_review",
commands=('python -c "print(\'should not run\')"',), agent="reviewer",
output="second-output.txt", on_status={"pass": "summarize"},
output="review.md",
), ),
StageConfig(id="third", type="summarize", output="final-notes.md"), StageConfig(id="implement", type="agent", agent="planner", output="impl.md"),
StageConfig(id="summarize", type="summarize", output="final-notes.md"),
) )
config = make_config(root, stages) config = make_config(root, stages)
runner = PipelineRunner(config, ArtifactStore(root, ".nightshift", run_id="test-run")) runner = PipelineRunner(config, ArtifactStore(root, ".nightshift", run_id="test-run"))
task = parse_tasks(TASK_MD)[0]
result = runner.run_task(parse_tasks(TASK_MD)[0]) result = runner.run_task(task)
task_dir = root / ".nightshift" / "runs" / "test-run" / "tasks" / "TASK-001"
self.assertEqual(result.status, "complete") self.assertEqual(result.status, "complete")
self.assertEqual([item.stage_id for item in result.stage_results], ["first", "third"]) self.assertEqual(result.retry_count, 0)
self.assertFalse((task_dir / "second-output.txt").exists()) self.assertEqual(
[r.stage_id for r in result.stage_results],
["plan", "review", "summarize"],
)
def test_on_status_routes_fail_to_target(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
_write_common_files(root)
fail_reviewer = 'python -c "print(\'status: fail\\nreason: bad plan\')"'
stages = (
StageConfig(id="plan", type="agent", agent="planner", output="plan.md"),
StageConfig(
id="review",
type="agent_review",
agent="reviewer",
on_status={"fail": "plan"},
output="review.md",
),
StageConfig(id="summarize", type="summarize", output="final-notes.md"),
)
config = make_config(root, stages)
config.agents["reviewer"] = AgentConfig(
id="reviewer",
backend="command",
command=fail_reviewer,
system_prompt=Path("reviewer.md"),
)
runner = PipelineRunner(config, ArtifactStore(root, ".nightshift", run_id="test-run"))
task = parse_tasks(TASK_MD)[0]
result = runner.run_task(task)
self.assertEqual(result.status, "failed")
self.assertEqual(result.retry_count, 2)
self.assertEqual(
[r.stage_id for r in result.stage_results],
["plan", "review", "plan", "review", "plan", "review"],
)
def test_on_status_escalate_routes_to_human_not_on_fail(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
_write_common_files(root)
escalate_reviewer = 'python -c "print(\'status: escalate\\nreason: need human\')"'
stages = (
StageConfig(id="plan", type="agent", agent="planner", output="plan.md"),
StageConfig(
id="review",
type="agent_review",
agent="reviewer",
on_status={
"retry": "plan",
"escalate": "human",
},
on_fail="plan",
output="review.md",
),
StageConfig(id="human", type="summarize", output="human-notes.md"),
StageConfig(id="summarize", type="summarize", output="final-notes.md"),
)
config = make_config(root, stages)
config.agents["reviewer"] = AgentConfig(
id="reviewer",
backend="command",
command=escalate_reviewer,
system_prompt=Path("reviewer.md"),
)
runner = PipelineRunner(config, ArtifactStore(root, ".nightshift", run_id="test-run"))
task = parse_tasks(TASK_MD)[0]
result = runner.run_task(task)
self.assertEqual(result.status, "complete")
self.assertEqual(result.retry_count, 1)
self.assertEqual(
[r.stage_id for r in result.stage_results],
["plan", "review", "human", "summarize"],
)
def test_on_fail_fallback_when_status_not_in_on_status(self) -> None:
with tempfile.TemporaryDirectory() as directory:
root = Path(directory)
_write_common_files(root)
fail_reviewer = 'python -c "print(\'status: fail\\nreason: bad\')"'
stages = (
StageConfig(id="plan", type="agent", agent="planner", output="plan.md"),
StageConfig(
id="review",
type="agent_review",
agent="reviewer",
on_status={"retry": "plan"},
on_fail="implement",
output="review.md",
),
StageConfig(id="implement", type="agent", agent="planner", output="impl.md"),
)
config = make_config(root, stages)
config.agents["reviewer"] = AgentConfig(
id="reviewer",
backend="command",
command=fail_reviewer,
system_prompt=Path("reviewer.md"),
)
runner = PipelineRunner(config, ArtifactStore(root, ".nightshift", run_id="test-run"))
task = parse_tasks(TASK_MD)[0]
result = runner.run_task(task)
self.assertEqual(result.status, "failed")
self.assertEqual(result.retry_count, 2)
self.assertEqual(
[r.stage_id for r in result.stage_results],
["plan", "review", "implement", "review", "implement", "review"],
)
def test_task_preflight_fails_when_task_specific_test_file_is_missing(self) -> None: def test_task_preflight_fails_when_task_specific_test_file_is_missing(self) -> None:
with tempfile.TemporaryDirectory() as directory: with tempfile.TemporaryDirectory() as directory: