bench-task-author

Benchmark task authoring and verification involves a step-by-step lifecycle process including source checks, redundancy avoidance, and predefined validation checks to ensure software testing tasks are high-quality and fair.

To start benchmark task creation, initiate the process using the 'bench-task-author start' command, which provides guided authorship and validation checks throughout the entire task lifecycle.

Benchmark task review works by supporting the revision of authored tasks against predefined quality standards and metrics, ensuring fairness and quality before tasks enter the evaluation bar.

Yes, you can use bench-task-author for verifying software testing tasks without external dependencies, as it facilitates verification by running tasks against a set of predefined checks and metrics.

Benchmark tasks need to pass verification checks to ensure they meet strict quality and fairness standards set by the task authoring persona before entering the evaluation bar.

During benchmark task authoring, limitations include the need to avoid source redundancy and the requirement to pass predefined validation checks, ensuring tasks meet quality standards before task verification.