handover-engine-controller

# Handover — Engine Controller Goal: preserve the running engine state so the next engine-controller can resume without guessing. ## Wind-down When Blake says wind down: - no new ratifications or widening unless Blake reverses; - do not spawn new Claude subagents except to preserve/finish already-active work; - let detached reviewer/codex continue only if reviewer-runner records PID/log/handoff; - write your own handover with `docs/worklog/new-note.sh --handover engine-controller`; - include active tasks, branches, worktrees, commits, subprocess/subagent PIDs, blockers, first next action; - create the next agent-comm thread only after all role handovers exist, then tell Blake the path. ## Rules - Stop widening: no new ratifications or task starts unless Blake reverses. - Do not spawn new Claude subagents during handover except to preserve/finish already-active work. - Do not kill active Claude subagents silently; they are attached to this session and may die with it. - No raw SQL, no broad cleanup, no unrelated staging. - Use the worklog script for the note; do not hand-name files. ## Create the note ```bash docs/worklog/new-note.sh --handover engine-controller ``` Read the printed path before editing. ## Fill only live state Include: - active thread path; - daemon/CLI health; - active tasks with status, branch, worktree, commit, next action; - active Claude subagents: purpose, parent/session risk, expected output, whether to wait; - deploy/accept/rebase blockers; - dirty worktrees/stashes that must not be dropped; - first exact next action for the next engine-controller. Do not paste SOP or long history. SOP belongs in skills; detailed history belongs in worklog/agent-comm. ## New thread After all three role handovers exist, engine-controller creates the next thread and posts links to the notes. Then wait for Blake to start the three next agents and say “start the engine.”