AgentSkillsCN

implementation-plan

在本地开发与 CI/CD 环境下生成一份确定性执行就绪的实施计划。在栈选型确认所选栈之后使用。

SKILL.md
--- frontmatter
name: implementation-plan
description: Generate a deterministic execution-ready implementation plan with local dev and CI/CD. Use after stack-selection confirms the chosen stack.

Implementation Plan Generator

Produce a deterministic, execution-ready implementation plan including local dev runbook and CI/CD setup. This skill does not execute – it only generates the plan.

Prerequisites

  • Selected stack from stack-selection
  • Architecture brief from architecture-refinement

Plan Phases

The implementation plan covers 10 mandatory phases:

  1. Foundation - Accounts, environments, regions, DNS
  2. Networking & Security - IAM, secrets, network boundaries
  3. Data Layer - Database provisioning, backups, migrations
  4. Local Development - Local orchestration, dependencies, parity
  5. CI/CD Pipeline - Build, test, deploy automation
  6. Application Delivery - Scaffolding, runtime config
  7. Observability - Logging, metrics, tracing, alerting
  8. Cost Controls - Budgets, guardrails, tagging
  9. Reliability - DR strategy, failover, RTO/RPO
  10. Hardening - Security hardening, runbooks

The following phases have dedicated standalone skills that can also be invoked directly:

  • local-dev-plan - Phase 4: Generate local development environment plans
    • "Create a local dev setup for React + PostgreSQL"
  • cicd-plan - Phase 5: Generate CI/CD pipeline configurations
    • "Create GitHub Actions workflow for my app"

Output Contract

yaml
implementation_plan:
  plan_id: "<unique id>"
  created_at: "<ISO timestamp>"
  stack_id: "<reference>"

  assumptions:
    - "<assumption>"

  phases:
    - phase: 1
      name: "Foundation"
      goals: []
      steps:
        - id: "1.1"
          name: "<step name>"
          description: "<description>"
          outputs: []

    - phase: 2
      name: "Networking & Security Baseline"
      goals: []
      steps: []

    - phase: 3
      name: "Data Layer Setup"
      goals: []
      steps: []

    - phase: 4
      name: "Local Development Environment"
      goals: []
      steps: []

    - phase: 5
      name: "CI/CD Pipeline"
      goals: []
      steps: []

    - phase: 6
      name: "Application Delivery"
      goals: []
      steps: []

    - phase: 7
      name: "Observability"
      goals: []
      steps: []

    - phase: 8
      name: "Cost Controls"
      goals: []
      steps: []

    - phase: 9
      name: "Reliability Posture"
      goals: []
      steps: []

    - phase: 10
      name: "Hardening Checklist"
      goals: []
      steps: []

  summary:
    total_phases: 10
    total_steps: <count>
    critical_decisions: []
    risks:
      - risk: "<risk>"
        mitigation: "<mitigation>"

Phase 1: Foundation

Goals: Set up cloud accounts, define environments, configure access

Steps:

  • 1.1 Create Cloud Accounts/Projects
  • 1.2 Define Environment Model (dev/staging/prod)
  • 1.3 Configure Regions
  • 1.4 Set Up DNS

Phase 2: Networking & Security

Goals: Implement least-privilege, secrets management, network security

Steps:

  • 2.1 Design IAM Model
  • 2.2 Configure Secrets Management
  • 2.3 Define Network Boundaries
  • 2.4 Apply Security Policies

Phase 3: Data Layer

Goals: Provision database, configure backups, establish migrations

Steps:

  • 3.1 Provision Database
  • 3.2 Configure Backups
  • 3.3 Set Up Migrations
  • 3.4 Configure HA/Replication (if required)

Phase 4: Local Development

Use the local-dev-plan skill for detailed planning (can be invoked directly).

Goals: Enable full local development, minimize time-to-first-run

Key outputs:

  • Local orchestration setup (docker-compose/k8s/hybrid/aspire)
  • Dependencies runbook
  • Seeding & migrations workflow
  • Local secrets strategy
  • Prod parity documentation

Offline Considerations (when required)

If offline requirement is transient or higher, include:

  • Offline Storage Strategy:

    • Technology selection (IndexedDB, SQLite WASM, localStorage)
    • Data model for offline access
    • Storage limits and quota handling
  • Sync Strategy:

    • Background sync implementation
    • Conflict resolution approach (last-write-wins, merge, CRDT)
    • Retry and backoff policies
  • Failure Modes:

    • Network detection and status handling
    • Graceful degradation patterns
    • User feedback during offline/sync states
  • Testing Offline Scenarios:

    • Offline mode simulation in local dev
    • Sync conflict test cases
    • Data integrity verification

Phase 5: CI/CD Pipeline

Use the cicd-plan skill for detailed planning (can be invoked directly).

Goals: Automate build/test/deploy, enforce quality gates

Key outputs:

  • Repository structure
  • Branching strategy
  • CI pipeline (build, test, lint, security scan)
  • CD pipeline per environment
  • Rollback procedures

Phase 6: Application Delivery

Goals: Scaffold application, configure runtime

Steps:

  • 6.1 Application Scaffolding
  • 6.2 Runtime Configuration
  • 6.3 Environment Variable Contract
  • 6.4 Deployment Topology

Phase 7: Observability

Goals: Centralized logging, metrics, tracing, alerting

Steps:

  • 7.1 Logging Setup
  • 7.2 Metrics Setup
  • 7.3 Dashboards
  • 7.4 Tracing Setup
  • 7.5 Alerting

Phase 8: Cost Controls

Goals: Budget visibility, guardrails, attribution

Steps:

  • 8.1 Budget Alerts
  • 8.2 Cost Guardrails
  • 8.3 Tagging Strategy

Phase 9: Reliability

Goals: Meet availability target, implement DR

Steps:

  • 9.1 DR Strategy
  • 9.2 Failover Configuration
  • 9.3 RTO/RPO Validation

Phase 10: Hardening

Goals: Security hardening, operational readiness

Steps:

  • 10.1 Security Hardening Checklist
  • 10.2 Operational Runbooks
  • 10.3 Operational Readiness Review

Next Step

This plan is ready for execution. Consider invoking skill-improvement at end of flow to capture learnings.