executing-plans

AI代理系统性地遵循实施计划,进行进度跟踪、质量关卡和阻塞问题解决。用于实施功能、遵循计划或执行任务时使用。

📖 中文导读:AI代理系统性地遵循实施计划,进行进度跟踪、质量关卡和阻塞问题解决。用于实施功能、遵循计划或执行任务时使用。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: executing-plans description: AI agent follows implementation plans systematically with progress tracking, quality gates, and blocker resolution. Use when implementing features, following plans, or executing tasks.

Executing Plans

Quick Start

  1. Prepare - Read entire plan, verify understanding, check dependencies
  2. Execute Task - Read -> Navigate -> Implement -> Verify -> Complete
  3. Track Progress - Update status after each task, report blockers immediately
  4. Pass Quality Gates - Validate after task, after phase, before completion
  5. Handle Deviations - Document changes, get approval for scope changes

Features

Feature Description Guide
Preparation Phase Understand before starting Read plan, verify deps, clarify questions
Task Execution Systematic per-task flow Read, navigate, implement, verify, complete
Progress Tracking Real-time status updates Completed X, In Progress Y, Remaining Z
Quality Gates Validation checkpoints Per-task, per-phase, pre-completion
Blocker Resolution Unblock systematically Document, assess, attempt, escalate, workaround
Deviation Handling Adapt to discoveries Document, assess scope/timeline, get approval

Common Patterns

# Task Execution Checklist
BEFORE STARTING:
[ ] Read task description completely
[ ] Understand acceptance criteria
[ ] Check dependencies completed
[ ] No blockers present

DURING IMPLEMENTATION:
[ ] Navigate to specified files
[ ] Understand existing code context
[ ] Implement changes incrementally
[ ] Write tests alongside code

AFTER COMPLETION:
[ ] All acceptance criteria met
[ ] Tests pass locally
[ ] No lint/type errors
[ ] Task marked complete

# Quality Gates
AFTER EACH TASK:
npm run lint && npm run typecheck && npm test -- --related

AFTER EACH PHASE:
npm run lint && npm run typecheck && npm test && npm run build

BEFORE COMPLETION:
npm run lint && npm run typecheck && npm test --coverage && npm run build
# Progress Report Template
## Status: [Date Time]

### Summary
- Completed: [N] tasks
- In Progress: [Task name]
- Remaining: [M] tasks
- Blockers: [None | Description]

### Current Task
**[Task Name]:** [Description]
- Started: [Time]
- Progress: ~60%
- Expected completion: [Time]

### ETA
On track for [date] | Delayed by [reason]
# Blocker Resolution Flow
BLOCKER DETECTED
      |
1. DOCUMENT immediately
   - What is blocked?
   - Why is it blocked?
      |
2. ASSESS impact
   - Critical path affected?
   - Time sensitive?
      |
3. ATTEMPT resolution (15-30 min)
   - Quick research
   - Alternative approach
      |
   +-> RESOLVED -> Continue
      |
4. ESCALATE if not resolved
   - Report with: issue, attempts, ask
      |
5. WORK AROUND while waiting
   - Move to non-blocked tasks

Best Practices

Do Avoid
Read entire plan before starting Skipping preparation phase
Verify understanding of each task Assuming task requirements
Execute tasks in sequence (unless parallel) Ignoring quality gates
Validate after each task completion Hiding blockers or delays
Report blockers immediately Deviating without documenting
Document deviations as they occur Rushing quality for speed
Complete fully before marking done Leaving tasks partially done
Ask questions early Skipping testing steps
  • writing-plans - Create plans to execute
  • verifying-before-completion - Quality gate checklists
  • thinking-sequentially - Track execution progress
  • solving-problems - Resolve blockers systematically