Configure token-bias controls on supporting model APIs for constrained generation, then validate the returned values.
Install
npx skills add alivirgo/Major-AI-Skills --skill logit-bias-token-steeringnpx skills add alivirgo/Major-AI-Skills --skill logit-bias-token-steering -a claude-code -gnpx skills add alivirgo/Major-AI-Skills --skill logit-bias-token-steering -a codex -ggemini skills install https://github.com/alivirgo/Major-AI-Skills.git --path skills/logit-bias-token-steeringgh skill install alivirgo/Major-AI-Skills logit-bias-token-steeringgit clone --depth 1 https://github.com/alivirgo/Major-AI-Skills /tmp/alivirgo__Major-AI-Skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/alivirgo__Major-AI-Skills/skills/logit-bias-token-steering ~/.claude/skills/logit-bias-token-steering # Claude Code
cp -R /tmp/alivirgo__Major-AI-Skills/skills/logit-bias-token-steering ~/.agents/skills/logit-bias-token-steering # Codex, Cursor, Gemini CLI, Copilot, OpenCode