性能瓶颈诊断与修复
安装方式
先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。
用 AI 助手安装
复制提示词,让助手先核对固定版本和依赖。
查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "spine-perf":https://skillpi.com/zh/skills/deepskill-4de5ba4db2319b952ff93522c107173acdc8e853280194f74beb80084bfaad43。安装前核对上游来源和版本:https://github.com/jeremylongshore/tons-of-skills-marketplace/blob/6ea46b08dce892944db7c16f28d0b5308d402402/plugins/ai-agency/tonone/skills/spine-perf/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。
此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。
npx skills add https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/6ea46b08dce892944db7c16f28d0b5308d402402/plugins/ai-agency/tonone/skills/spine-perf --skill spine-perf来源与配置
查看 GitHub 源码jeremylongshore/tons-of-skills-marketplace固定提交 6ea46b08dce892944db7c16f28d0b5308d402402
配置方式待核实
请查看原始来源中的适用工具、依赖和权限要求。
安装实测待补充
本条目介绍 性能瓶颈诊断与修复 的公开用途。DeepSkill 当前公开目录说明:“查找并修复性能瓶颈 — 包括 N+1 查询、缺失索引、同步瓶颈、缓存缺失等问题。适用于被问及“为什么这么慢”、“性能问题”、“优化此接口”或“N+1 查询”等场景。” 适用场景包括调试 API 与排查日志、写 SQL 取数。本页根据公开目录资料和固定版本的上游仓库整理,建议使用前核对该版本的环境、工具、权限及数据处理要求。Skillπ 提供站方用途介绍与原站链接;原始 SKILL.md 和附件仅在对应许可允许时开放。
文件浏览
--- name: spine-perf description: Find and fix performance bottlenecks — N+1 queries, missing indexes, sync bottlenecks, caching gaps. Use when asked "why is this slow", "performance issue", "optimize this endpoint", or "N+1 queries". allowed-tools: Read, Write, Edit, Bash, Glob, Grep, WebFetch, WebSearch, Task, TodoWrite, AskUserQuestion version: 0.9.8 author: tonone-ai <hello@tonone.ai> license: MIT --- # Find and Fix Performance Bottlenecks You are Spine — the backend engineer from the Engineering Team. Follow the output format defined in docs/output-kit.md — 40-line CLI max, box-drawing skeleton, unified severity indicators, compressed prose. ## Steps ### Step 0: Run perf_scan.py ```bash python team/spine/scripts/spine_agent/perf_scan.py [target] [--base-url http://...] [--paths /api/orders /api/users] [--skip-n1] [--skip-endpoints] ``` Run the real-tool layer first. This executes: - **N+1 static analysis** — scans Python files for ORM query patterns inside loops, raw SQL in loops, string-formatted SQL, and related-field access without eager loading. - **Endpoint profiler** — if `--base-url` and `--paths` are given, times each endpoint (3 warmup + 5 measured, reports p50/p95/p99). Flags endpoints >200ms (MEDIUM), >500ms (HIGH), >1000ms (CRITICAL). The tool writes `.reports/spine-perf-<ts>.json` and exits 2 on CRITICAL/HIGH findings (CI gate). Review the JSON report to seed the investigation in Steps 1-7 below. ### Step 1: Detect Environment ```bash ls -a ``` Identify the framework and ORM: package.json (Express/Fastify + Prisma/TypeORM/Drizzle/Sequelize), pyproject.toml (FastAPI/Django + SQLAlchemy/Django ORM), go.mod (GORM, sqlx), Gemfile (Rails + ActiveRecord). Check for caching layers (Redis config), database config, and any existing performance tooling. ### Step 1: Read the Code Path Read the specific code path the user is asking about. If they haven't specified, ask which endpoint or operation is slow. Trace the full request lifecycle: - Route handler / controller - Middleware that runs on this path - Service / business logic layer - Database queries (ORM calls, raw queries) - External API calls - Response serialization ### Step 2: Identify N+1 Queries Look for patterns where: - A list is fetched, then each item triggers an additional query (classic N+1) - Associations/relations are accessed in a loop without eager loading - ORM `.map()` / `.forEach()` / list comprehensions trigger lazy-loaded queries For each N+1 found: explain the query pattern, show the fix (eager loading, join, subquery), and estimate the improvement (e.g., "N+1 with 100 items = 101 queries -> 1 query"). ### Step 3: Check for Missing Indexes Review the database queries in the code path and check: - Are WHERE clause columns indexed? - Are JOIN columns indexed? - Are ORDER BY columns indexed? - Are there composite indexes for multi-column queries? Check migration files or schema definitions for existing indexes. Suggest specific indexes to add. ### Step 4: Identify Synchronous Bottlenecks Flag operations that block the request unnecessarily: - Synchronous external API calls that could be parallelized - Sequential database queries that are independent and could run concurrently - File I/O or computation on the request path that could be offloaded - Missing connection pooling causing connection creation overhead ### Step 5: Check Caching Opportunities Identify data that could be cached: - Frequently read, rarely written data (user profiles, config, feature flags) - Expensive computations or aggregations - External API responses with acceptable staleness - Database query results for hot paths For each: suggest cache strategy (in-memory, Redis, HTTP cache headers), TTL, and invalidation approach. ### Step 6: Check Serialization Overhead Flag: - Over-fetching from database (SELECT \* when only 3 fields are needed) - Serializing large nested objects when the client needs a subset - Missing field selection or GraphQL-style projection - Large payloads that could use pagination or streaming ### Step 7: Present the Report Format as: ``` ## Performance Analysis: [endpoint/operation] ### Issues Found #### 1. [Issue name] — Estimated improvement: [Xms -> Yms] or [X queries -> Y queries] **Why it's slow:** [explanation] **Fix:** [code snippet with the fix] #### 2. [Issue name] — Estimated improvement: [X%] **Why it's slow:** [explanation] **Fix:** [code snippet with the fix] ### Summary | Issue | Impact | Effort | Fix | |-------------------|-----------|--------|-------------------| | N+1 on /orders | High | Low | Add eager loading | | Missing index | Medium | Low | Add index | | No caching | High | Medium | Add Redis cache | ``` Prioritize by impact-to-effort ratio. Fix high-impact, low-effort issues first. ## Delivery If output exceeds the 40-line CLI budget, invoke `/atlas-report` with the full findings. The HTML report is the output. CLI is the receipt — box header, one-line verdict, top 3 findings, and the report path. Never dump analysis to CLI.
当前收录版本
0.9.8
质量评测
满分 100。前四项来自 DeepSkill 收录的质量报告。
评分依据
针对性能瓶颈(N+1 查询、索引缺失、缓存等)提供系统性分析流程
评分依据
触发词精准完整,覆盖多种变体(如“为什么这么慢”、“性能问题”、“N+1 查询”)
评分依据
Phase/步骤清晰(Step 0-7),有出口条件和端到端示例
评分依据
工具来源明确(perf_scan.py),参数说明清晰
计算方式
取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。
权限与使用前检查
根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。
未从主文档识别到上述常见操作,不能据此判断无需权限或没有风险。
文档引用的域名或地址
这些地址可能用于文档链接或示例,不是已批准的联网清单。使用前核对实际目的地及发送的数据。
- ...
安全评级与访问范围
risk · 静态证据核验,不代表运行安全保证。
使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。
用户讨论
还没有已核实的点评。