Work Next Command

November 12, 2025 · View on GitHub

Usage: /sk:work-next

Description: Get intelligent recommendations for the next work item to start based on dependencies and priority.

Overview

The work-next command analyzes your work items and recommends the most important work that can be started immediately. It considers:

  • Dependency satisfaction (only shows unblocked items)
  • Priority levels (critical > high > medium > low)
  • Work item types and context
  • Current project state

Usage

/sk:work-next

No arguments needed - the command automatically analyzes all work items.

Recommendation Algorithm

The command follows this logic:

1. Check for Urgent Items (Highest Priority)

  • ⚠️ Urgent items override all other logic
  • If an urgent item exists with status not_started, it's returned immediately
  • Urgent items ignore dependencies and ignore priority levels
  • Only ONE work item can be urgent at a time

2. Filter Available Work Items

  • Status must be not_started
  • All dependencies must be completed
  • Excludes in_progress, blocked, and completed items

3. Sort by Priority

  • critical - Blocking issues (highest priority)
  • high - Important work
  • medium - Normal priority
  • low - Nice to have (lowest priority)

4. Recommend Top Item

  • Returns the highest priority unblocked work item
  • Provides rationale for the recommendation
  • Shows context of other waiting items

Output Format

RECOMMENDATION
==============

Next Work Item: feature_auth
Title: Add user authentication
Type: feature | Priority: high

RATIONALE:
- Highest priority unblocked work item
- All dependencies satisfied:
  ✓ feature_database (completed)
  ✓ refactor_models (completed)

CONTEXT:
Ready to Start (3 items):
  🟠 feature_auth: Add user authentication
  🟡 refactor_api: Refactor API error handling
  🟢 feature_dashboard: Add admin dashboard

Blocked (2 items):
  🟠 feature_search: Waiting on feature_auth
  🟡 integration_test_api: Waiting on feature_auth, bug_timeout

In Progress (1 item):
  🔴 bug_timeout: Fix database timeout

QUICK START:
  /sk:start feature_auth

Examples

Standard Recommendation

/sk:work-next

Output:

RECOMMENDATION
==============

Next Work Item: bug_session_timeout
Title: Fix session timeout error
Type: bug | Priority: critical

RATIONALE:
- Critical priority (blocks other work)
- All dependencies satisfied (no dependencies)
- Urgent bug fix needed

CONTEXT:
Ready to Start (4 items):
  🔴 bug_session_timeout: Fix session timeout (RECOMMENDED)
  🟠 feature_auth: Add user authentication
  🟡 refactor_api: Refactor API error handling
  🟢 feature_dashboard: Add admin dashboard

Blocked (0 items)

QUICK START:
  /sk:start bug_session_timeout

Urgent Item (Overrides Everything)

/sk:work-next

Output:

⚠️  URGENT ITEM DETECTED

ID: bug_production_hotfix
Title: Critical production hotfix
Type: bug
Priority: medium

This item requires immediate attention and overrides normal priority.
To start: /start bug_production_hotfix

Note: When an urgent item exists:

  • It's returned immediately regardless of priority level
  • It ignores dependencies (can be returned even if dependencies are incomplete)
  • All other work items are deprioritized until the urgent item is completed
  • Only ONE work item can be urgent at a time

Multiple High-Priority Items

/sk:work-next

Output:

RECOMMENDATION
==============

Next Work Item: feature_auth
Title: Add user authentication
Type: feature | Priority: high

RATIONALE:
- Tied for highest priority (high) with 2 other items
- Selected: feature_auth (created first)
- Blocks 3 other work items (high impact)

CONTEXT:
Ready to Start (3 items):
  🟠 feature_auth: Add user authentication (RECOMMENDED)
  🟠 feature_payment: Implement payment system
  🟠 feature_notifications: Add email notifications

Blocked (3 items):
  🟠 feature_search: Waiting on feature_auth
  🟡 feature_dashboard: Waiting on feature_auth
  🟡 integration_test_api: Waiting on feature_auth

NOTE: feature_auth blocks 3 other items. Completing it will
      unblock significant downstream work.

QUICK START:
  /sk:start feature_auth

No Ready Items

/sk:work-next

Output:

NO RECOMMENDATION
=================

No work items are ready to start.

Blocked (3 items):
  🟠 feature_search: Waiting on feature_auth (in_progress)
  🟡 refactor_api: Waiting on feature_auth (in_progress)
  🟢 feature_dashboard: Waiting on feature_auth, feature_search (both incomplete)

In Progress (1 item):
  🟠 feature_auth: Add user authentication

NEXT STEPS:
1. Complete feature_auth to unblock 3 dependent items
2. Use /sk:end to complete current work
3. Create new work items: /sk:work-new

All Work Completed

/sk:work-next

Output:

NO RECOMMENDATION
=================

All work items are completed! 🎉

Summary:
  ✓ 15 work items completed
  ✓ 0 work items remaining

NEXT STEPS:
- Create new work items: /sk:work-new
- Review completed work: /sk:work-list --status completed
- Plan next milestone: /sk:work-list --milestone next_phase

Understanding the Context Section

Ready to Start

Work items with:

  • Status: not_started
  • All dependencies: completed
  • Can be started immediately

Sorted by priority (critical → high → medium → low)

Blocked

Work items waiting on dependencies:

  • Status: blocked or not_started with incomplete dependencies
  • Shows which dependencies are blocking
  • Cannot start until dependencies complete

In Progress

Currently active work items:

  • Status: in_progress
  • Being worked on in current or recent session
  • Should be completed before starting new work

Priority Indicators

  • 🔴 critical - Blocking issue, urgent requirement
  • 🟠 high - Important work to be done soon
  • 🟡 medium - Normal priority work
  • 🟢 low - Nice to have, can be deferred

Impact Analysis

The command identifies high-impact work items:

Blocks multiple items:

NOTE: feature_auth blocks 5 other items. Completing it will
      unblock significant downstream work.

Critical path item:

NOTE: feature_database is on the critical path. It determines
      the earliest project completion date.

Milestone critical:

NOTE: This item is required for milestone "Phase 1 MVP" (due soon).

Integration with Other Commands

After Getting Recommendation

/sk:work-next
# See recommendation: feature_auth

/sk:start feature_auth
# Start working on recommended item

Compare with Manual Selection

/sk:work-next
# Algorithm recommendation

/sk:work-list --status not_started
# See all available items

/sk:start <chosen_item>
# Start your choice

Check After Completing Work

/sk:end
# Complete current work item

/sk:work-next
# Get next recommendation automatically

Recommendation Strategies

Strategy 1: Priority-First (Default)

Always picks highest priority unblocked item

  • Best for: Projects with clear priorities
  • Pro: Ensures important work gets done first
  • Con: May ignore high-impact low-priority items

Strategy 2: Impact-Aware

Consider both priority and downstream impact

  • The command shows impact in rationale
  • You can manually choose high-impact items
  • Look for "blocks N items" notes

Strategy 3: Milestone-Driven

Focus on items in current milestone

  • Use /sk:work-list --milestone current_sprint
  • Filter by milestone first
  • Then apply priority sorting

When to Override Recommendation

Consider starting a different item when:

  1. Context Switching Cost

    • Recommended item requires different skillset
    • Current mental context suits different item better
  2. External Dependencies

    • Waiting on stakeholder feedback
    • Third-party API access needed
  3. Team Coordination

    • Another team member working on related code
    • Avoiding merge conflicts
  4. Time Constraints

    • Quick wins needed before deadline
    • Low-priority but fast items available

Error Handling

No Work Items Exist

/sk:work-next

Output:

NO RECOMMENDATION
=================

No work items found.

Create your first work item:
  /sk:work-new

All Items In Progress

/sk:work-next

Output:

NO RECOMMENDATION
=================

All work items are either completed or in progress.

In Progress (2 items):
  🟠 feature_auth
  🔴 bug_timeout

Complete current work before starting new items:
  /sk:end

Best Practices

1. Check Regularly

# After completing work
/sk:end
/sk:work-next

# Before starting new work
/sk:work-next
/sk:start

2. Review Context

  • Don't just start recommended item blindly
  • Review blocked items and their dependencies
  • Consider downstream impact

3. Update Priorities

# If recommendation doesn't match needs
/sk:work-update <item_id> priority
# Then check again
/sk:work-next

4. Break Down Large Items

If recommended item is too large:

# Create smaller sub-items
/sk:work-new  # Create smaller task
/sk:work-update <new_item> add-dependency
# Add large item as dependency

See Also