Claude Code Plugins

Community-maintained marketplace

Feedback

>-

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name executing-tasks
description Use when working through any task checklist (not just OpenSpec). Complete one task, verify, mark done, then next. No skipping, no adding, no reordering.

Executing Tasks

Work through task lists systematically.

Announce: "I'm using executing-tasks to work through this checklist systematically."

Iron Law

ONE TASK AT A TIME - VERIFY BEFORE MOVING ON

Process

Step 1: Identify Current Task

Find the first incomplete task:

- [x] Completed task
- [x] Completed task
- [ ] ← THIS ONE (current)
- [ ] Pending task

Step 2: Announce

Working on: [Task description]

Step 3: Execute

Do EXACTLY what the task says:

  • If it says "create file X" → create file X
  • If it says "add function Y" → add function Y
  • If it says "run command Z" → run command Z

Do NOT:

  • Add things the task didn't ask for
  • Skip steps you think are unnecessary
  • Combine with other tasks
  • "Improve" the approach

Step 4: Verify

Run any verification specified in the task.

If no verification specified, at minimum:

  • Did I do what the task asked?
  • Did I break anything else?

Step 5: Mark Complete

Update the task list: - [ ]- [x]

Step 6: Report

Completed: [Task description]
Result: [Brief outcome]
Next: [Next task] or "All tasks complete"

Step 7: Repeat

Go to Step 1 for next task.

Handling Blockers

If a task cannot be completed:

BLOCKED: [Task description]
Reason: [Why it can't be done]
Options:
1. [Possible resolution]
2. [Alternative approach]
3. Skip and continue (if independent)

Awaiting guidance.

Do NOT:

  • Guess at the solution
  • Skip without reporting
  • Change the task scope

Task Dependencies

If tasks have dependencies:

- [ ] 1.1 Create database table
- [ ] 1.2 Add RLS policy (depends on 1.1)
- [ ] 2.1 Create frontend component (independent)
  • Complete dependencies first
  • Independent tasks can be done in any order
  • If blocked by dependency, report it

Progress Tracking

Keep a running count:

Progress: 3/10 tasks complete
Current: Task 1.4

Completion

When all tasks are [x]:

All tasks complete: [X/X]

Summary:
- [Category]: [What was done]
- [Category]: [What was done]

Verification:
- [What was verified]

Ready for: [Next step - usually review]

Red Flags - STOP

If you catch yourself:

  • Working on multiple tasks at once
  • Skipping "obvious" tasks
  • Adding unrequested work
  • Not verifying before marking complete

STOP. One task. Verify. Mark complete. Next.