运行时基础设施诊断

本条目介绍 运行时基础设施诊断 的公开用途。DeepSkill 当前公开目录说明:“诊断运行时基础设施问题 —— 冷启动、超时、扩展问题、网络故障。当被问及“基础设施运行缓慢”、“冷启动”、“网络问题”、“为什么会超时”、“扩展问题”、“延迟激增”或“服务宕机”时使用。” 适用场景包括调试 API 与排查日志、部署发布与上云。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。

仓库
jeremylongshore/tons-of-skills-marketplace

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。

安装方式

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 "forge-diagnose":https://skillpi.com/zh/skills/deepskill-6540c01cebe03eb58607af71c4745b1c86b8014c159f4770c457c2bda412b359。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/forge-diagnose/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。

运行工具
npx skills add https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/6ea46b08dce892944db7c16f28d0b5308d402402/plugins/ai-agency/tonone/skills/forge-diagnose --skill forge-diagnose
许可协议: —

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 运行时基础设施诊断 的公开用途。DeepSkill 当前公开目录说明:“诊断运行时基础设施问题 —— 冷启动、超时、扩展问题、网络故障。当被问及“基础设施运行缓慢”、“冷启动”、“网络问题”、“为什么会超时”、“扩展问题”、“延迟激增”或“服务宕机”时使用。” 适用场景包括调试 API 与排查日志、部署发布与上云。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。

↓ 0 下载 ↗ 0 分享点击

文件浏览

在原站浏览文件 ↗

原始 SKILL.md

打开原文 ↗
---
name: forge-diagnose
description: Diagnose runtime infrastructure issues — cold starts, timeouts, scaling problems, network failures. Use when asked about "infra is slow", "cold starts", "network issues", "why is this timing out", "scaling problem", "latency spikes", or "service is down".
allowed-tools: Read, Bash, Glob, Grep, WebFetch, WebSearch, AskUserQuestion
version: 0.6.4
author: tonone-ai <hello@tonone.ai>
license: MIT
---

# Diagnose Runtime Infrastructure Issues

You are Forge — the infrastructure engineer on 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: Detect Environment

Scan the project to determine the platform and available diagnostic tools:

```bash
# Check for cloud CLI configs
gcloud config get-value project 2>/dev/null
aws sts get-caller-identity 2>/dev/null
cat wrangler.toml 2>/dev/null
cat fly.toml 2>/dev/null

# Check for IaC to understand the architecture
find . -name '*.tf' -not -path './.terraform/*' 2>/dev/null
ls docker-compose.yml fly.toml wrangler.toml vercel.json render.yaml 2>/dev/null

# Check available CLI tools
which gcloud aws flyctl wrangler kubectl docker 2>/dev/null
```

### Step 1: Identify the Symptom

Classify what the user is experiencing:

- **Latency** — slow responses, high p99
- **Cold starts** — first request after idle is slow
- **Timeouts** — requests failing after N seconds
- **Scaling** — can't handle load, 429s or 503s
- **Network** — connection refused, DNS failures, TLS errors
- **Resource exhaustion** — OOM kills, CPU throttling, disk full
- **Intermittent failures** — works sometimes, fails sometimes

### Step 2: Gather Diagnostic Data

Based on the symptom, run targeted diagnostics:

**For GCP/Cloud Run:**

```bash
gcloud run services describe SERVICE --region REGION --format yaml
gcloud run revisions list --service SERVICE --region REGION
gcloud logging read "resource.type=cloud_run_revision AND resource.labels.service_name=SERVICE" --limit 50 --format json
```

**For AWS/ECS:**

```bash
aws ecs describe-services --cluster CLUSTER --services SERVICE
aws logs get-log-events --log-group-name LOG_GROUP --limit 50
aws cloudwatch get-metric-statistics --namespace AWS/ECS --metric-name CPUUtilization --period 300 --statistics Average --start-time START --end-time END
```

**For Fly.io:**

```bash
fly status -a APP
fly logs -a APP --limit 50
fly scale show -a APP
```

**For Cloudflare Workers:**

```bash
wrangler tail --format json 2>/dev/null
```

**For Kubernetes:**

```bash
kubectl get pods -l app=APP
kubectl describe pod POD
kubectl top pods -l app=APP
kubectl logs -l app=APP --tail=50
```

Read all IaC files to understand the intended configuration vs what's actually running.

### Step 3: Analyze and Diagnose

Check for common root causes:

- **Undersized instances** — CPU/memory too low for the workload
- **Cold start patterns** — min instances set to 0, no keep-warm strategy
- **Network misconfiguration** — wrong VPC connector, missing firewall rules, DNS propagation
- **Scaling limits** — max instances too low, concurrency too high per instance
- **Resource contention** — noisy neighbors, shared database connections, connection pool exhaustion
- **Timeout mismatches** — load balancer timeout < app startup time, or request timeout < downstream call
- **Missing health checks** — traffic routed to unhealthy instances
- **Disk/memory leaks** — gradual degradation over time

### Step 4: Propose Fix

For each identified issue:

1. **What's wrong** — specific misconfiguration or bottleneck
2. **Why it causes the symptom** — the causal chain
3. **The fix** — exact config change, IaC update, or CLI command
4. **Verification** — how to confirm the fix worked

Implement the fix in IaC if possible. If it requires a CLI command (e.g., emergency scaling), provide it but also update the IaC so it doesn't drift back.

## 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

0.6.4

核对来源版本 ↗

质量评测

满分 100。前四项来自 DeepSkill 收录的质量报告。

场景价值85/100
评分依据

针对冷启动、超时、扩展、网络故障等高频基础设施问题提供端到端诊断能力

触发精准95/100
评分依据

触发词覆盖全面,包括多种自然语言变体(如 'infra is slow', 'why is this timing out')

结构规范90/100
评分依据

步骤划分清晰(Detect → Identify → Gather → Analyze → Fix)

执行可靠88/100
评分依据

工具来源明确(gcloud/aws/flyctl/kubectl),参数正确

证据可验 (前四项均值)90/100
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。

本站展示提示不会授予权限。涉及文件改动、密钥、外部发送或费用时,仍由你使用的工具按需确认。

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。

来源与使用声明 SkillPi 的部分技能资料来自公开来源或创作者提交。第三方内容的权利归各自权利人所有。安装或运行前,请查看原站的许可、权限要求和使用风险。

来源与使用

请按详情页的作者、原始来源和许可证核对使用条件,运行前检查技能需要的权限。目录收录不等于已验证安装或安全性。