run-qectorlisted

# Run QECTOR You operate the QECTOR Workbench headless controller. Every number is traced to a real decode - never fabricate a benchmark result, LER value, or export. ## Verified surface (wheel 1.0.0) All symbols below were verified against the published wheel on this device (module listing, signature inspection, and live runs). The entire Workbench surface is **provisional / non-frozen** (changelog 0.7.0 -> 1.0.0, 1.0.0 API freeze note) - never quote it as the stable library contract. ```python from qector_decoder_v3.workbench import Workbench # WorkbenchError, Job also exported wb = Workbench() ``` | Method | Verified behavior | |---|---| | `submit_job(spec: dict) -> str` | Queues a benchmark; returns job_id. Single FIFO worker thread. | | `wait(job_id: str, timeout: float | None = None) -> dict` | Blocks until done; returns the job status dict. | | `get_job(job_id: str) -> dict` | Job status: `status`, `progress`, `units_done`, `units_total`, `has_artifact`, `spec`, `error`, `submitted_unix`, `started_unix`, `finished_unix`, `job_id`. | | `job_artifact(job_id: str) -> dict | None` | The artifact dict: `{"environment", "spec", "results"}`. | | `list_jobs() -> list[dict]` | All known jobs. | | `cancel_job(job_id: str) -> str` | Cooperative cancel (queued jobs cancel instantly; running jobs stop at the next decoder/distance unit boundary). | | `run_benchmark(spec: dict, job: Job | None = None) -> dict` | Synchronous benchmark; same artifact shape. | | `load_stim(source