brewcodespeclisted

<instructions> ## Input Handling | Input | Action | |-------|--------| | `$ARGUMENTS` empty | Read `.claude/TASK.md` → first line = path → derive task dir | | `$ARGUMENTS` has text | Use as task description | | `$ARGUMENTS` has path | Read file as task description | ### Flag Parsing Parse `$ARGUMENTS` for flags BEFORE input detection: | Flag | Effect | |------|--------| | `-n`, `--noask` | Skip all user questions, auto-approve defaults | Strip flag from `$ARGUMENTS`. Remaining text = description or path. ### Naming - Timestamp: `YYYYMMDD_HHMMSS` (e.g., `20260208_143052`) - Name slug: lowercase, underscores, from description (e.g., `auth_feature`) - Task dir: `.claude/tasks/{TIMESTAMP}_{NAME}_task/` ### Workflow 0. **Check Adapted Templates** (REQUIRED FIRST) **EXECUTE** using Bash tool: ```bash test -f .claude/tasks/templates/SPEC.md.template && echo "SPEC.md.template" || echo "SPEC.md.template MISSING" ``` > **STOP if MISSING** — Run `/brewcode:setup` first. 1. **Read & Analyze Input** - Parse `$ARGUMENTS` per Input Handling table - Determine scope: files affected, areas of codebase - Identify what needs clarification 2. **Clarifying Questions** (AskUserQuestion) **If `--noask`:** Skip. Record in SPEC User Q&A: "Skipped (--noask mode)". Infer scope from description and codebase analysis. **Otherwise:** Use AskUserQuestion tool to ask 3-5 questions, grouped in batches of up to 4 per AskUserQuestion call. Focus on: | # | Categ