← ClaudeAtlas

nextlisted

Use when someone asks what to do next in a BrotherSBE project. Evaluates a fixed priority ladder against observable state and returns exactly one recommended action with a one sentence reason, never a menu of options. Invoke as /brothersbe:next.
khalilmaaouni/BrotherSBE · ★ 0 · Web & Frontend · score 70
Install: claude install-skill khalilmaaouni/BrotherSBE
# Next One question, one answer. This skill exists so the user never has to hold the lifecycle in their head. ## LANE C1 (B-003): one canonical next action `src/brothersbe/lifecycle.py`'s `reduce_next_action` is now the ONE place that decides which outstanding fact is most urgent for a change. Plain `sbe status --json`'s `nextAction` (a sentence) and `nextActionDetail` (`{actionId, label, reason, basis}`), and `sbe status --team --json`'s own severity-10 finding for each discovered change, are both derived through that SAME reducer, so they can no longer read the same recorded state two different ways. This skill's own job shrank to match: read that field and recommend exactly what it says, rather than re-deriving a priority ladder by hand from the rest of the JSON. Do not re-implement the ladder here; if it ever needs to change, it changes in `lifecycle.py`, once, for every reader. ## Rungs 1-3: checks the reducer cannot run `sbe status` never starts a subprocess and never runs a NEW check over source code (its own module docstring names this as the reason a truthful summary stops and says so rather than run one); it reads state other commands already recorded. These three rungs are genuinely outside what `nextAction` can ever cover for exactly that reason, so they stay separate, live probes. Evaluate them, in this order, BEFORE trusting anything `nextAction` says, and stop at the first one that matches: 1. **Environment broken.** Run `"${CLAUDE_PLUGIN_ROOT}/bin/sbe" d