用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/mikailustuner/OmniRule --skill internal-platforms命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
基于 SOC 职业分类
| name | internal-platforms |
| description | Internal Platforms: Self-service infrastructure, golden paths, platform engineering. |
| triggers | {"extensions":[".yaml",".json",".go"],"directories":["platform/","internal/","infrastructure/"],"keywords":["platform","internal","self-service","golden path","developer platform","idp","paved road"]} |
| auto_load_when | Building internal developer platforms or self-service infrastructure |
| agent | platform-engineer |
| tools | ["Read","Write","Bash"] |
Focus: Self-service, golden paths, platform engineering
Platform Building Blocks:
├── Compute
│ ├── Self-service Kubernetes clusters
│ ├── Serverless functions
│ └── Container registry
│
├── Data
│ ├── Database provisioning (managed)
│ ├── Cache (Redis) provisioning
│ └── Data pipeline templates
│
├── Networking
│ ├── VPC per team automatic
│ ├── DNS management
│ └── SSL/TLS certificates
│
├── Observability
│ ├── Pre-configured logging
│ ├── Metrics dashboards
│ └── Alerts (service-level)
│
└── CI/CD
├── Shared pipeline templates
├── Environment provisioning
└── Deployment automation
Self-Service Design:
├── Portal/API-first
│ ├── Web UI for humans
│ └── CLI/API for automation
│ └── Terraform provider for tools
│
├── Approval workflows
│ ├── Auto-approve safe requests
│ ├── Escalation for risky changes
│ └── RBAC on requests
│
├── Guardrails
│ ├── Prevents misconfigurations
│ ├── Enforces security policies
│ └── Defaults to safe
│
└── Cost tracking
└── Show cost per team/project
└── Budget alerts
└── Chargeback (optional)
Golden Path Principles:
├── Pre-configured templates
│ ├── Docker images
│ └── Helm charts
│ └── Pipeline configs
│
├── Opinionated defaults
│ ├── Standardize choices
│ └── Reduce decision fatigue
│ └── Best practices built-in
│
├── Clear boundaries
│ ├── What teams can customize
│ └── What is fixed
│ └── Examples: "You can add env vars, not change language"
│
└── Documentation
└── How-to guides
└── Examples
└── Do/don't
Platform Team Responsibilities:
├── Build platform
│ ├── Infrastructure as code
│ └── Self-service tooling
│ └── Golden paths
│
├── Operate platform
│ ├── Reliability (SLOs)
│ └── Incident response
│ └── Capacity planning
│
├── Enable users
│ ├── Developer onboarding
│ └── Support (tickets)
│ └── Training
│
└── Gather feedback
└── Regular user surveys
└── Usage metrics
└── Product backlog
Platform Metrics:
├── Adoption
│ ├── % of services on platform
│ └── % of deployments via self-service
│ └── Daily active users
│
├── Velocity
│ ├── Time to first deployment
│ ├── Time to provision resource
│ └── Lead time for changes
│
├── Reliability
│ ├── Platform uptime
│ └── Incident count
│ └── MTTR (mean time to recovery)
│
└── Developer experience
├── NPS (platform team)
└── Ticket response time
└── Documentation usage
❌ Platform for platform's sake — no real users
✅ Solve specific developer pain points
❌ Too many choices — analysis paralysis
✅ Opinionated defaults, limited options
❌ No onboarding — steep learning curve
✅ Templates, docs, support
❌ No SLAs — unreliable for users
✅ Define SLO, communicate when broken
❌ Build alone — ignore feedback
✅ Regular user feedback loops
| Component | Tool | Purpose |
|---|---|---|
| K8s platform | Rancher, Anthos | Self-service K8s |
| Provisioning | Terraform, Crossplane | Infrastructure |
| Service catalog | Backstage, Port | Discovery |
| CI/CD | Tekton, ArgoCD | Pipelines |
| Monitoring | Prometheus, Grafana | Observability |