Aa import

aa-import

Resolve the Agent-Agile playbook root: use the first of these that exists — (1) ${CLAUDE_PLUGIN_ROOT}/playbooks, (2) ./.claude/agent-agile/playbooks, (3) ./.agents/agent-agile/playbooks, (4) ~/.claude/agent-agile/playbooks, (5) ~/.agents/agent-agile/playbooks, (6) ./playbooks.

Read playbooks/planner.md section Import path in full, and follow it exactly — do NOT re-derive or improvise the system. That section also depends on The 8 intake questions, Pushback patterns, and Drafting the plan; read those too before starting.

Running the import

The skill's argument is a path to, or description of, the user's existing plan document.

1. Read the document in full — no skimming.
2. Answer the 8 intake questions from the document alone, without asking the user anything yet. Mark any thinly-answered question as a gap rather than guessing, using the same "thin answer" bar defined in The 8 intake questions.
3. Ask the user only the gaps, one question at a time, with the same pushback discipline as a fresh intake.
4. Deliver the split statement exactly as defined at the end of planner.md's The 8 intake questions — the "my part / your part" narration plus the process preview — before writing any files. Imported plans get it too: their authors are the users most likely to assume the system does everything.
5. Map the result onto the plan format and check the three things Import path names explicitly: whether the Key Results pass the three KR tests, whether every epic has a demo sentence, and whether any existing stories pass the worker-readiness test.
6. Write .planning/PROJECT.md, .planning/ROADMAP.md, and Epic 1's story cards from the templates at playbooks/templates/PROJECT.md, playbooks/templates/ROADMAP.md, and playbooks/templates/STORY.md, plus .planning/PREREQS.md and .planning/CONFIG.md from their templates. Produce a gap list alongside these files — what was missing or broken in the source document — not a silent rewrite. On a multi-epic import, a missing risk register is itself a gap to report: the source plan named no domain edges and no fixtures (see playbooks/planner.md "Research fan-out and the risk register"), and the user decides whether to run the fan-out before executing.
7. Update .planning/STATE.md from playbooks/templates/STATE.md, Phase intake.

Panel and exits

Instruct the user to run /aa-panel next — an imported plan gets the same adversarial review as a freshly drafted one, no leniency for "it was already written." After the panel returns, resolve to exactly one of the three exits the playbook defines:

• Proceed — translate the document into the plan format and run the current epic.
• Reshape — back to intake, carrying the panel's findings as the starting gaps.
• Kill — write .planning/LEARNINGS.md from playbooks/templates/LEARNINGS.md explaining why, and stop.