plan

# Plan Skill

Create structured implementation plans from goals, research, or feature requests.
Produces beads issues with proper dependencies and wave computation for parallel execution.

## Overview

**Core Purpose**: Transform a goal into an executable plan with discrete beads issues,
dependency ordering, and wave-based parallelization for `/crank` (autonomous) or `/implement-wave` (supervised).

**Key Capabilities**:
- 6-tier context discovery hierarchy
- Prior plan discovery to prevent duplicates
- Feature decomposition with dependency modeling
- Beads issue creation with epic-child relationships
- Wave computation for parallel execution

**When to Use**: Work needs 2+ discrete issues with dependencies.
**When NOT to Use**: Single task (use `/implement`), exploratory (use `/research`).

---

## Instructions

### Phase 0: Rig Detection

**CRITICAL**: All `.agents/` artifacts go to `~/gt/.agents/<rig>/` based on the primary codebase being planned.

**Detection Logic**:
1. Identify which rig's code you're planning (e.g., files in `~/gt/ai-platform/` → `ai-platform`)
2. If planning across multiple rigs, use `_cross-rig`
3. If unknown/unclear, ask user

| Files Being Read | Target Rig | Output Base |
|------------------|------------|-------------|
| `~/gt/athena/**` | `athena` | `~/gt/.agents/athena/` |
| `~/gt/hephaestus/**` | `hephaestus` | `~/gt/.agents/hephaestus/` |
| `~/gt/daedalus/**` | `daedalus` | `~/gt/.agents/daedalus/` |
| Multiple rigs | `_cross-rig` | `~/gt/.agents/_cross-rig/` |

```bash
# Set RIG variable for use in output paths
RIG="athena"  # or hephaestus, daedalus, _cross-rig
mkdir -p ~/gt/.agents/$RIG/plans/
```

---

### Phase 1: Context Discovery

See `core-kit/skills/research/references/context-discovery.md` for full 6-tier hierarchy.

**Quick version**: Code-Map → Semantic Search → Scoped Grep → Source → .agents/ → External

**Checklist**:
- [ ] Checked code-map index
- [ ] Ran semantic search (if MCP available)
- [ ] Followed signposts to s