“runbook”

Database Migration Runbook Writer

Turns a schema change into a safe zero-downtime runbook: expand/contract steps, locks analysis, verification, rollback.

Incident On-Call Runbook Generator

Writes actionable runbooks for common alerts: triage decision tree, commands, mitigation before diagnosis, escalation thresholds.

SRE On-call Runbook：故障排查与应急响应指南

为SRE on-call工程师提供全面的运行手册，涵盖常见系统问题、详细故障排除步骤、明确的升级流程及关键监控指标，确保高效响应与问题解决…

Alerts: SLO-based burn-rate alerts (multi-window), symptoms over causes,

响应

Operability: metrics/alerts/SLO, feature flags, rollback plan, launch plan,

Response metrics

Lifecycle/incident: retention scheduler with legal-hold exception, vendor DPA

准备

事后活动

Preparation

Post-incident activity

Agentic RAG代码与工程知识检索

构建由智能体驱动的检索增强系统，支持查询规划、多跳检索、工具化读取与带引用 综合，面向代码库、文档与工单的工程知识问答…

Incident Response Playbook Agent

构建事件响应剧本智能体，在事件指挥官授权下引导处置已确认事件（含AI专项事件）；自动动作可逆且留痕，外部联络由人工发起。

Agentic RAG for Code & Engineering Knowledge

构建由智能体驱动的检索增强系统，支持查询规划、多跳检索、工具化读取与带引用综合，面向代码库、文档与工单的工程知识问答…

Python Production Script / CLI Generator

Generates a robust Python CLI with argparse, logging, retries, dry-run, and typed config — ideal for automation, data pipelines, and internal ops scripts.

Microservice / Event-Driven System Coder

Scaffolds an event-driven service with outbox, schema registry, contracts, and observability — keeps services loosely coupled.

CI/CD Pipeline Engineer

Creates secure, fast pipelines: caching, matrix tests, SBOM, signing, progressive delivery, and rollback.

Production-Grade MCP Server in TypeScript

面向企业内网的 TypeScript 版 MCP 服务器工程模板，同时支持 stdio 与 Streamable HTTP 双传输，用于把内部系统安全、可审计地开放给 AI 编程助手。

Multi-Tenant MCP Gateway with Auth, Rate Limiting and Audit

企业级多租户 MCP 统一网关，集中处理 OAuth 鉴权、限流、参数过滤、双人审批与全量审计，将数十个上游 MCP 服务收敛为单一受管入口。