AgentSkillsCN

post-plan-setup

运行完整的计划后流程：精炼计划 → 创建 Linear 问题 → 设置 CLAUDE.md 文件。在 /plan-project 生成 v1 计划后使用此功能。各阶段之间可暂停，以供选修式的审阅。

SKILL.md
--- frontmatter
name: post-plan-setup
user-invocable: true
description: >
  Runs the full post-plan workflow: refine plan → create Linear issues →
  setup CLAUDE.md. Use after /plan-project produces a v1 plan. Pauses
  between phases for optional review.
allowed-tools:
  - mcp__sequential-thinking__sequentialthinking
  - mcp__linear__*
  - Read
  - Write
  - Bash(find:*)
  - Bash(cat:*)
  - Bash(ls:*)
  - Glob
  - Grep
agent: post-plan-orchestrator

Instructions

You are orchestrating the full post-plan setup workflow. This runs three phases in sequence, pausing between each for human review.

Input

  • •The v1 plan file path is passed as $ARGUMENTS
  • •If no argument provided, look for the most recent plan file in the project root or docs/ directory

Phase 1: Refine Plan

  1. •Invoke the /refine-plan skill with the v1 plan
  2. •Wait for completion
  3. •Present a summary of the refined plan to the user:
    • •Total number of tasks created
    • •Task dependency overview
    • •Any flags or ambiguities found
  4. •Ask the user: "The refined plan is ready at docs/project-plan-refined.md. Would you like to review and make changes before I create Linear issues, or should I continue?"
  5. •If the user wants changes, apply them and re-validate
  6. •If the user approves (or says continue), proceed to Phase 2

Phase 2: Create Linear Issues

  1. •Invoke the /create-issues skill with the refined plan
  2. •Wait for completion
  3. •Present a summary to the user:
    • •Number of issues created
    • •Link to the Linear project/view
    • •Any issues that needed special handling
  4. •Ask the user: "All Linear issues are created. Would you like to review them before I set up CLAUDE.md, or should I continue?"
  5. •If the user wants changes, apply them via the Linear MCP
  6. •If the user approves, proceed to Phase 3

Phase 3: Setup CLAUDE.md

  1. •Invoke the /setup-claude-md skill
  2. •Wait for completion
  3. •Present the generated CLAUDE.md to the user for review
  4. •Ask the user: "CLAUDE.md is ready. Would you like any changes?"
  5. •Apply any requested changes

Completion

Once all three phases are done, present a final summary:

  • •Refined plan location and task count
  • •Linear issues created with project link
  • •CLAUDE.md location and section overview
  • •Suggested next step: "Pick up Task 1 from the refined plan or assign issues to your team"

Error Handling

  • •If any phase fails validation after retries, stop the workflow
  • •Present what succeeded and what failed
  • •Suggest the user run the failed skill individually with /refine-plan, /create-issues, or /setup-claude-md after fixing the issue