TasksCodingEngineering track

Dependency-ordered task runner

eng-code-04suite v1.0code

Quality25%judge panel
Correctness60%deterministic · hidden-tests
Completeness15%judge panel

Exact prompt sent to every model

Write runner.py with async def run(tasks: dict[str, Task], *, concurrency: int = 4) -> Report. Task is a dataclass with deps: list[str] and fn: Callable[[], Awaitable[Any]]. Run tasks in dependency order with at most concurrency running at once. If a task raises, mark it failed, mark every transitive dependent skipped, and keep running everything else. Detect cycles before starting and raise CycleError naming the cycle. Detect unknown dependency names and raise KeyError. Report is a dataclass with succeeded, failed (name to exception), skipped (name to the failed ancestor name), and order (start order). Use asyncio only. Type hints throughout. Return only the file in one code block.

Responses

Latest published run: gemini-3.8-flash. For each model the repeat closest to its published median is shown; every repeat is in the repository.

No published attempts for this task yet

When the run is published, every model's raw response, per-dimension scores, judge rationales, human review notes and rendered artifacts appear here side by side.