AgentSkillsCN

2-implement

按照既定计划执行任务,并从上次中断的地方继续推进。

SKILL.md
--- frontmatter
name: 2-implement
description: Execute the plan and continue where you left off
argument-hint: "{planname}[/NN] [focus or step]"

Focus

$ARGUMENTS


Execute the plan step-by-step, or continue where you left off.

Plan Loading

Parse $ARGUMENTS to find plan:

  • {planname} → Load plans/{planname}/scope.md
  • {planname}/NN → Load plans/{planname}/NN_*.md (e.g., myplan/01)
  • If no planname, list existing plans and ask

Read the plan file before starting. Check [x] markers to find where to resume.

Resuming Work

  1. Check progress markers ([x] vs [ ]) in plan file
  2. Find next incomplete step
  3. If sub-plan complete, prompt to move to next sub-plan
  4. If argument specifies step, start from there

Core Principles

Follow the Plan

  • Work through the plan sequentially
  • Do not implement tests here (that's /5-test)
  • Any deviation requires discussion - STOP and discuss before continuing

Deviation Tracking

When you must deviate:

  1. STOP and explain why
  2. Get user confirmation
  3. Document in plans/{planname}/deviation_{scope|NN}.md

Progress Tracking

After completing each step:

  • Mark the step as [x] in the plan file

Simplicity

  • Write the simplest code that handles the full complex case
  • Avoid clever abstractions - prefer straightforward logic
  • If code is hard to read, it's wrong

Break Things Up

  • Split files when they exceed ~200 lines
  • Split functions when they exceed ~30 lines

What Belongs Here vs Later

  • Here: Core logic, data structures, primary happy path
  • Later (/4-quality): Design review
  • Later (/5-test): Test coverage
  • Don't gold-plate. Get working code first.

Anti-Patterns

  • No stubs, TODOs, or placeholders
  • No try/catch unless actually handling the error
  • No copy-paste with minor modifications

If you encounter blockers, stop and discuss.