Pick Next Plan Shell

Pick the next draft plan shell from .turbo/plans/ by reading YAML frontmatter, then carry it through the planning pipeline: expand → refine → self-improve → halt.

Task Tracking

At the start, use TaskCreate to create a task for each step:

1. Scan shells and pick next
2. Run /expand-plan-shell skill
3. Run /refine-plan skill
4. Run /self-improve skill
5. Halt with next-step instructions

Step 1: Scan Shells and Pick Next

Glob .turbo/plans/*.md and read each file's YAML frontmatter. Filter to files with type: shell. Categorize each shell:

• Draft — status: draft. Candidate if all depends_on entries have status: done.
• Ready — status: ready. Already expanded. Re-enter the pipeline at Step 3 (refine).
• In-progress — status: in-progress. Implementation started. Suggest running /implement-plan with its path and stop.
• Done — status: done. Completed. Skip.

Priority order for picking:

1. Any shell with status: ready — re-enter at Step 3
2. Any shell with status: draft whose depends_on are all done — start at Step 2
3. Nothing actionable — see terminal conditions below

If multiple candidates exist at the same priority level, pick the one with the lowest shell number (from the NN in the filename). If ambiguous, use AskUserQuestion to let the user choose.

State the picked shell path, its status, and dependencies before continuing.

If no shell files exist, tell the user to run /turboplan for a new task (which routes to /draft-spec + /draft-plan-shells for complex projects) and stop.

If all shells are done, report completion:

All shells are done. The project is complete.

If remaining shells are blocked, report which shells are blocked and which dependencies they're waiting on.

Step 2: Run /expand-plan-shell Skill

Run the /expand-plan-shell skill, passing the shell file path. /expand-plan-shell sets the shell's status to ready on completion.

Step 3: Run /refine-plan Skill

Run the /refine-plan skill, passing the expanded plan path. Loops until the plan stabilizes.

Step 4: Run /self-improve Skill

Run the /self-improve skill to compound planning learnings.

Step 5: Halt with Next-Step Instructions

Halt with this message:

Plan ready at <plan path>.

Planning context is likely full, and the plan is comprehensive enough to continue fresh. Run /clear, then /implement-plan <slug> to implement. After that, run /pick-next-plan-shell again for the next shell.

Rules

• For draft shells, all pipeline steps run. Ready shells re-enter at Step 3, skipping Step 2.
• Do not edit plan files directly. Revisions go through /refine-plan or /expand-plan-shell.
• Never modify the spec file.
• Do not pre-verify Consumes or refresh surveys here. /expand-plan-shell handles that.
• Do not attempt to auto-implement. The user drives implementation with /implement-plan in a fresh session.
• If a shell file is missing or has invalid frontmatter, halt and report.