fix(ralph-loop): keep LLM in smart zone with fresh context per iteration#1348

Open

edxeth wants to merge 7 commits intocode-yeongyu:devfrom

Commits on Feb 12, 2026

feat(ralph-loop): add context_strategy with 'reset' default for fresh context per iteration
edxeth
committed
feat(ralph-loop): add --strategy flag and fix todo-continuation conflict
edxeth
committed
docs: update ralph-loop documentation with --strategy flag and default_strategy config
edxeth
committed
fix: normalize --strategy flag to lowercase for case-insensitive matching
edxeth
committed
fix(ralph-loop): ensure reset strategy sends identical prompt as iteration 1
edxeth
committed
fix(ralph-loop): let OpenCode handle iteration 1 session naming
edxeth
committed
chore: regenerate schema to match origin/dev
edxeth
committed