What problem does it solve? Turning a raw problem idea into a complete, contest-ready competitive programming package requires coordinating many error-prone steps: choosing OI vs ICPC format, designing subtask scoring, writing model and brute-force solutions, stress testing, building testlib validators and generators, and auditing the result. This Skill sequences that entire pipeline with gated checkpoints so nothing ships half-finished. ## Core Features & Use Cases - Gated pipeline orchestration: Drives the full flow from idea to package — shaping, coding, stress testing, test generation, attack zoo, statement, and review — delegating each phase to sibling skills while enforcing order and loop-backs. - Machine-readable progress tracking: Resumes work from disk via tools/package_status.py and problem.json rather than chat memory, so interrupted sessions continue at the first incomplete phase. - OI/ICPC format handling: Blocks on format questions, enforces subtask ladders with point splits summing to 100 for OI, and applies kill-test policies appropriate to each format. - Use Case: A problem setter says "take this idea to a Polygon package" — the Skill asks the format questions, produces problem.json, writes sol-main.cpp and sol-brute.cpp, stress tests them, generates testlib validators and generators, attacks the suite with a WA/TLE zoo, and finishes with a reviewed package. ## Quick Start Ask the assistant to take your problem idea and intended solution all the way to a Polygon-ready package, mentioning whether you want OI or ICPC format.
