forge-sprintlisted

# forge-sprint — a sprint view over the task DAG `/forge:sprint` slices the project's task DAG into bounded **sprints**: `plan` commits the next ready tasks to a sprint file; `review` reports what got done vs. what carries over. It is **not** a parallel tracker — it references **T-IDs only**, and the task DAG (`pipeline/05-plan/task-dag.md`) plus implementation progress (`pipeline/06-implementation/progress.md`) remain the single source of truth (F-044/045). The selection is deterministic — `scripts/sprint.py` does it (no LLM); this skill invokes it and reports the result. ## When to Use - `/forge:sprint plan` — commit the next chunk of ready tasks to `sprint-NN.md`. - `/forge:sprint review` — at sprint end, summarize done vs. carried into a review file. - `/forge:sprint list` — show existing sprints and their progress. ## When NOT to Use - You just want to build a milestone end-to-end → `/forge:build --milestone N`. - There is no task DAG yet → run `/forge:plan` (Stage 5) first. ## Steps 1. **Plan a sprint.** ```bash python3 ${CLAUDE_PLUGIN_ROOT}/scripts/sprint.py plan --cwd . \ [--size K] [--milestone N] ``` It selects the next `K` (default 5) **ready** tasks in dependency order — **carry-over first**: any task committed to the previous sprint that is not yet done in `progress.md` leads the new sprint and keeps its T-ID (F-047) — then writes `pipeline/05-plan/sprint-NN.md` and prints the committed T-IDs to stdout. Show the user the sprin