| name | next |
| description | Recommend what issue to work on next based on priority, complexity, and dependencies. Use when user asks 'what should I work on', 'what's next', or 'pick something for me'. |
| user-invocable | true |
| argument-hint | [optional: 'quick' for low complexity, 'important' for P1 only] |
Recommend Next Issue
Analyze the backlog and recommend the best issue to work on next.
Workflow
1. Fetch Open Issues and Board Status
gh issue list --state open --limit 100 --json number,title,labels,body
gh project item-list 1 --owner eshaffer321 --format json --limit 100 | jq '[.items[] | select(.status == "In Progress") | .content.number]'
If the gh project command is rate-limited, skip it and just recommend based on issue labels/priority. Note to the user that you couldn't check the board for in-progress items.
Important:
- Exclude issues that are already "In Progress" on the project board (if available) - another agent may be working on them.
- Exclude issues with the
epic label - these are parent issues tracked via sub-issues, not worked on directly.
2. Apply Recommendation Logic
Priority Order:
- P0-critical (always first)
- P1-high
- P2-medium
- P3-low
Within Same Priority, Consider:
- Complexity (prefer low complexity for quick wins)
- Dependencies (issues that unblock other work)
- Area clustering (if user recently worked on combat, suggest related combat issues)
3. Quick Win Filter
If user says /next quick or asks for "something quick":
- Filter to
complexity:low only
- Prioritize P1 > P2 within that set
4. Important Filter
If user says /next important:
- Filter to
P0-critical and P1-high only
- Include even high-complexity items
5. Present Recommendations
Provide top 3 recommendations with reasoning:
## Recommended: #32 Mandatory Card Play/Discard Per Turn
**Why this one:**
- P1-high priority (important for core rules)
- Low complexity (can complete quickly)
- No dependencies on other issues
- Foundational rule that should be enforced early
**Labels:** P1-high, complexity:low, area:turn, area:cards, edge-case
**Summary:** The engine doesn't enforce that players must play at least one card per turn...
---
## Alternatives
**#33 Resting as State** (P1-high, complexity:high)
- Also high priority but more complex
- Good if you have more time
**#58 Unit Combat Integration** (P1-high)
- Important for gameplay
- Moderate scope
6. Offer to Start
After recommendation, offer:
"Would you like me to start working on #32? Just say 'yes' or /work #32"
Consideration Factors
- Momentum: If user just finished a combat issue, suggest related combat work
- Balance: If lots of edge-cases done recently, suggest a feature
- Dependencies: Note if an issue depends on another (rare, but check issue body for "depends on" or "blocked by")
- Time of day: If late, suggest quick wins over complex tasks (if known)
Example
User: /next
Response:
"Based on your backlog, I recommend #32: Mandatory Card Play/Discard Per Turn
- Priority: P1-high
- Complexity: Low
- Area: Turn management, cards
- Why: This is a fundamental rule enforcement that's quick to implement. It will improve game accuracy without a large time investment.
Alternatives:
- #33 Resting State (P1-high but complex)
- #58 Unit Combat Integration (P1-high, broader scope)
Want me to start on #32?"