running-contests

Automating competitive programming contest execution requires orchestrating problem retrieval, iterative code implementation, and verdict verification. This skill manages the full submission loop autonomously while tracking progress to ensure full contest completion across multiple problems in a single session.

Yes, you can run Codeforces contests in ICPC scoring mode. The skill adheres to specific scoring models like ICPC or IOI, binding to judge-specific MCP tools to fetch problem statements and manage submissions for platforms such as Codeforces and AtCoder.

The best way to handle algorithm design and submission loops is through an orchestrated iterative process. This skill manages the cycle of algorithm design, C++ code implementation, and automated verdict polling, maintaining momentum while providing checkpoints for ambiguous problem statements.

Yes, you need judge-specific MCP tools to participate in automated competitive programming. The skill binds to these MCP tools to perform platform operations like fetching problem statements and submitting code, acting as the orchestrator for the end-to-end contest lifecycle.

Contest automation handles stuck problems or errors using robust error handling and hybrid autonomy. It operates autonomously to maintain momentum but provides clear checkpoints for ambiguous statements or difficult problems, ensuring progress tracking continues throughout the session.

The running-contests skill is judge-agnostic and supports platforms like AtCoder and Codeforces. It binds to any judge MCP to perform operations such as fetching problem statements and submitting code, allowing flexible contest execution across different competitive programming environments.