compare-porting-feasibility

Automate porting feasibility evaluations by orchestrating extraction and reporting across multiple evals simultaneously. The workflow reads an eval list, spawns Task agents for parallel processing, and maps eval names to directory slugs to coordinate assessments.

Pairwise comparisons generate an HTML dashboard after all individual assessments complete. The system coordinates these comparisons across evals and produces the dashboard to visualize the porting feasibility results.

Run the pipeline with the --report-only option to generate per-eval reports without repeating the extraction phase. This skips YAML assessment file generation and focuses solely on producing the reporting output.

YAML assessment files are written during the extraction phase to capture porting feasibility data for each eval. These structured files serve as the foundation for generating per-eval reports and running pairwise comparisons.

Manual mapping is unnecessary because the orchestration automatically maps eval names to directory slugs. This automated mapping ensures correct file paths when reading eval lists and writing assessment outputs.

Use --force to overwrite existing assessment files and reports during the porting workflow. This ensures all evals are re-processed and updated outputs are generated, even if previous assessment results already exist.