원클릭으로
exploration-plan
Structure work into plans and Beads. Three modes: plan from brainstorm, plan from scratch, or create Beads from existing plan.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Structure work into plans and Beads. Three modes: plan from brainstorm, plan from scratch, or create Beads from existing plan.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | exploration-plan |
| description | Structure work into plans and Beads. Three modes: plan from brainstorm, plan from scratch, or create Beads from existing plan. |
Turn ideas into executable plans, and plans into Beads. This skill handles the full journey from exploration to tracked work items.
Which planning mode?
• Plan from brainstorm → reads ideas/<topic>/spec.md
• Plan from scratch → start with just an idea
• Create Beads from plan → reads planned/<topic>/plan.md
Input: exploration/ideas/YYYY-MM-DD-<topic>/spec.md
Output: plan.md in same folder, then move to planned/
exploration/ideas/spec.md not found, check for legacy exploration.mdideas/<topic>/ → planned/<topic>/If user says yes to Beads, execute Mode 3.
Input: User describes idea
Output: planned/YYYY-MM-DD-<topic>/plan.md
exploration/planned/YYYY-MM-DD-<topic>/If user says yes to Beads, execute Mode 3.
Input: exploration/planned/YYYY-MM-DD-<topic>/plan.md
Output: Beads (feature + tasks), move folder to mission/staged/
List available plans in exploration/planned/
Confirm selection with user
Read plan.md - Parse the feature and tasks
Show what will be created - List feature name + task names
Create Beads:
# Get active epic
EPIC_ID=$(bd list -t epic --status open --json | jq -r '.[0].id')
# Create feature under epic
bd create "Feature title" -t feature --parent "$EPIC_ID" -p 2
# Get the feature ID just created
FEATURE_ID=$(bd list -t feature --status open --json | jq -r '.[-1].id')
# Create tasks under feature with descriptions from plan.md
bd create "Task 1" -t task --parent "$FEATURE_ID" -p 1 -d "$(cat <<'EOF'
**Goal:** [One sentence goal from plan]
**Files:**
- [file list from plan]
**Steps:**
1. [steps from plan]
**Tests:**
- [ ] [test criteria from plan]
EOF
)"
bd create "Task 2" -t task --parent "$FEATURE_ID" -p 2 -d "$(cat <<'EOF'
**Goal:** [One sentence goal from plan]
**Files:**
- [file list from plan]
**Steps:**
1. [steps from plan]
**Tests:**
- [ ] [test criteria from plan]
EOF
)"
# Set up dependencies if specified in plan
bd dep add <task-id> <depends-on-id>
bd sync
Description format: Extract from each ### Task: section in plan.md:
**Goal:** from task's Goal line**Files:** from task's Files line**Steps:** from task's Steps list**Tests:** from task's Tests checklistRename and move folder - Use the feature number (e.g., 1.5 from space-agents-1.5) instead of the date: planned/YYYY-MM-DD-<topic>/ → mission/staged/<feature-num>-<topic>/
Confirm - "Feature ready. Run /mission to begin execution."
Advisors, not decision makers. HOUSTON synthesizes and can override.
space-agents:plan-task-planner - Breaks feature into tasksspace-agents:plan-sequencer - Analyzes dependencies, execution orderspace-agents:plan-implementer - Creates TDD task breakdown per taskSpawn all 3 sequentially (NOT as background tasks). Wait for each agent to complete and return results before spawning the next. HOUSTON must have all council input before synthesizing.
Use explicit hierarchy that maps to Beads:
# Feature: [Feature Name]
**Goal:** [One sentence description]
## Overview
[Context, motivation, what this achieves]
## Tasks
### Task: [Task 1 Name]
**Goal:** [One sentence]
**Files:** [Create/Modify/Test]
**Depends on:** [None or other task names]
**Steps:**
1. [Step description]
2. [Step description]
**Tests:**
- [ ] [Observable behaviour or outcome to verify]
- [ ] [Another verifiable criterion]
### Task: [Task 2 Name]
**Goal:** [One sentence]
**Files:** [Create/Modify/Test]
**Depends on:** Task 1 Name
**Steps:**
1. [Step description]
2. [Step description]
**Tests:**
- [ ] [Observable behaviour or outcome to verify]
- [ ] [Another verifiable criterion]
## Sequence
1. Task 1 (no dependencies)
2. Task 2 (depends on Task 1)
3. Task 3 (can run parallel with Task 2)
## Success Criteria
- [ ] [Criterion 1]
- [ ] [Criterion 2]
Parsing rules:
# Feature: → Creates Beads feature### Task: → Creates Beads task**Depends on:** → Sets up bd dep add**Tests:** → Included in task description for Builder/Inspectorexploration/
ideas/ ← /brainstorm creates spec.md here
planned/ ← /plan creates plan.md, moves from ideas/
mission/
staged/ ← /plan creates Beads, moves from exploration/planned/
complete/ ← /land moves here when feature closes
Execute a single task with Pathfinder/Builder/Inspector crew. Self-fetches work from Beads.
Select feature, select mode, delegate to execution skill.
Start a Space-Agents session. Displays welcome screen with project status.
HOUSTON spawns Pathfinder, Builder, Inspector per task. Best for medium features (4-10 tasks).
Launch ralph.sh in background. Lightweight mode (default) or full Pod crew (--pod). Best for large features (10+ tasks).
Explore ideas through conversation. HOUSTON asks questions, has opinions, and suggests background agents when investigation would help.