Set up coder-eval in this repository — scan for what is worth evaluating (Claude Code skills, an MCP server, a CLI), then scaffold a task directory with one real task and the command to run it.
Install
npx skills add UiPath/coder_eval --skill initnpx skills add UiPath/coder_eval --skill init -a claude-code -gnpx skills add UiPath/coder_eval --skill init -a codex -ggemini skills install https://github.com/UiPath/coder_eval.git --path plugins/coder-eval/skills/initgh skill install UiPath/coder_eval initgit clone --depth 1 https://github.com/UiPath/coder_eval /tmp/UiPath__coder_eval
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/UiPath__coder_eval/plugins/coder-eval/skills/init ~/.claude/skills/init # Claude Code
cp -R /tmp/UiPath__coder_eval/plugins/coder-eval/skills/init ~/.agents/skills/init # Codex, Cursor, Gemini CLI, Copilot, OpenCode