plan-expeditelisted

# Plan Expedite `/plan-expedite --plan <path>` runs the full plan -> sync -> handoff pipeline as one autonomous step. Default output: TWO continue commands in order — first a `/goal "<condition>"` line that arms the Stop hook over the agent-completable span, then the `/build-phase --plan <path>` command — both to run **in the same window** — no forced `/compact`, because auto-compaction handles context on its own when it fills and the `SessionStart` re-inject hook reloads `current.md` afterward. (A focused `/compact` before the long build is optional — see step 4.) Add `--new-window` to hand off to a fresh window instead: a durable `current.md` write (`task-handoff --next-task`), then `/session-wrap --end` renders the handoff to disk (`.claude/task-state/handoff-prompt.md`) and prints the Pick-up-here block. --- ## Execution model — autonomous, invoke sub-skills via the Skill tool (HEAVY) This skill's whole reason to exist is that the operator does not want to type `/plan-review` → wait → `/plan-wrap` → wait → `/repo-sync` → wait → `/session-wrap`. They invoked `/plan-expedite` to have those four happen as one autonomous run. Therefore: 1. **Execute, do not advise.** When invoked, you MUST invoke each sub-skill via the `Skill` tool in the order specified in the "Sub-skill chain" section. Do NOT respond by emitting the chain as text (e.g. "Next: `/plan-review` → `/plan-wrap` → `/repo-sync` → `/session-wrap` → `/build-phase`"). That listing-the-steps response is the single