plan-epicslisted

# /plan-epics — briefs → GitHub epics + user stories Turn the brief(s) in **$ARGUMENTS** into a tracked backlog: one or more **epics** (labelled `epic`), each broken into **user stories** (labelled `user-story` + area tags), created as GitHub issues and cross-linked. When the work spans **multiple epics, a `product-manager` subagent authors each epic's stories in parallel** — one PM per epic. Input (**$ARGUMENTS**): the brief(s). May be inline text, a path to a file/dir of briefs, or several briefs separated by `---` or numbered. If it's empty, ask the user for the brief before doing anything. --- ## Config (defaults — override only if the repo clearly needs it) - `REPO` = current repo (`gh repo view --json nameWithOwner -q .nameWithOwner`). - `EPIC_LABEL` = `epic`; `STORY_LABEL` = `user-story` (create if missing). - `AREA_LABELS` = derived from the repo's existing `gh label list` (prefer the `area:*` family). Create a new area label only when a story clearly needs one that doesn't exist — and say which you created. - Issue-body trailer: `Claude-Session: ...` on every epic and story (read from the harness/session context; do not invent). - `DRY_RUN` = on if the caller says "dry run" / "preview": print the full plan and create NOTHING. --- ## Stage 0 — Intake & context (orchestrator) 1. Parse `$ARGUMENTS` into one or more briefs; read any referenced files/dirs. 2. Gather repo context so the backlog fits the project: `README` / `CLAUDE.md` (domain, conventions,