Token导航 LogoToken导航TokenDH.com
研究检索执行命令github未标认证来源可访问许可证需确认审计通过

docker-opsDocker OPS 搜索

Agent Skill

用于辅助云资源、部署、容器、基础设施和运维自动化任务。它适合让 Agent 检查配置、整理部署步骤、分析资源状态、生成排障思路或辅助云服务接入。使用时需要明确目标环境、账号权限、区域和资源组,区分本地测试与生产操作;涉及删除资源、重启服务、修改网络或权限配置时,应先确认影响范围。

总安装

367

周安装

15

GitHub Stars

17

下载量

118
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:docker-ops(Docker OPS 搜索)
来源仓库:https://github.com/0xdarkmatter/claude-mods
仓库路径:skills/docker-ops
安装命令:
npx skills add https://github.com/0xdarkmatter/claude-mods --skill docker-ops
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/0xdarkmatter/claude-mods --skill docker-ops

简介

专注于 Docker 容器化应用的全生命周期管理, 涵盖构建、运行与编排。docker-ops 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

  • 适用于开发测试环境搭建、生产部署流程梳理及容器故障排查场景。
  • 通过提供标准化操作模板和安全实践建议提升自动化效率。
  • 需配合宿主环境权限使用,涉及关键系统操作时应二次确认影响范围。

SKILL.md

Docker Operations

Comprehensive Docker patterns for building, running, and composing containerized applications.

Dockerfile Best Practices

PracticeDoDon't
Base imageFROM node:20-slimFROM node:latest
Layer cachingCopy dependency files first, then sourceCOPY.. before RUN install
Package installapt-get update && apt-get install -y... && rm -rf /var/lib/apt/lists/*Separate RUN for update and install
UserUSER nonroot (create if needed)Run as root in production
Multi-stageSeparate build and runtime stagesShip compiler toolchains
Secrets--mount=type=secret (BuildKit)COPY.env. or ARG PASSWORD
ENTRYPOINT vs CMDENTRYPOINT for fixed binary, CMD for defaultsRelying on shell form for signal handling
WORKDIRWORKDIR /appRUN cd /app &&...
.dockerignoreInclude .git, node_modules, __pycache__No.dockerignore at all
LabelsLABEL org.opencontainers.image.*No metadata

Multi-Stage Build Decision Tree

Choose your runtime base image by language:

Go ──────────── CGO disabled? ──── Yes ──► scratch or distroless/static
                                   No ───► distroless/base or alpine

Rust ─────────── Static musl? ──── Yes ──► scratch or distroless/static
                                   No ───► distroless/cc or debian-slim

Node.js ──────── Need native? ──── Yes ──► node:20-slim
                                   No ───► node:20-alpine (smaller)

Python ────────── Need C libs? ─── Yes ──► python:3.12-slim
                                   No ───► python:3.12-slim (still slim)

Java ──────────── JRE only ──────────────► eclipse-temurin:21-jre-alpine
See: references/multi-stage-builds.md for complete annotated examples per language.

Layer Caching Rules

Docker caches each layer. A cache miss at layer N invalidates all subsequent layers.

What Invalidates Cache

TriggerEffect
Changed file in COPY/ADDInvalidates this layer + all below
Changed RUN command textInvalidates this layer + all below
Changed ARG valueInvalidates from the ARG declaration down
--no-cache flagInvalidates everything
Base image updateInvalidates everything

Optimal Layer Order

# 1. Base image (changes rarely)
FROM python:3.12-slim

# 2. System dependencies (changes rarely)
RUN apt-get update && apt-get install -y --no-install-recommends \
    libpq-dev \
    && rm -rf /var/lib/apt/lists/*

# 3. Dependency files (changes occasionally)
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# 4. Application code (changes frequently)
COPY src/ ./src/

# 5. Runtime config (changes frequently)
CMD ["python", "-m", "app"]

Rule of thumb: Order layers from least-frequently-changed to most-frequently-changed.

.dockerignore Essentials

# Version control
.git
.gitignore

# Dependencies (rebuilt in container)
node_modules
__pycache__
*.pyc
.venv
vendor/

# Build artifacts
dist/
build/
target/
*.egg-info

# IDE and editor
.vscode
.idea
*.swp
*.swo

# Docker files (prevent recursive context)
Dockerfile*
docker-compose*
.dockerignore

# Environment and secrets
.env
.env.*
*.pem
*.key

# Documentation and tests (unless needed)
docs/
tests/
*.md
LICENSE

Why it matters: Without .dockerignore, docker build sends the entire context directory to the daemon. A .git folder alone can add hundreds of megabytes.

Docker Compose Quick Reference

Service Definition

services:
  web:
    build:
      context: .
      dockerfile: Dockerfile
      target: production        # Multi-stage target
    image: myapp:latest
    ports:
      - "8080:8000"
    environment:
      DATABASE_URL: postgres://db:5432/app
    env_file:
      - .env
    volumes:
      - ./src:/app/src          # Bind mount (dev)
      - app-data:/app/data      # Named volume (persistent)
    depends_on:
      db:
        condition: service_healthy
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 10s
    restart: unless-stopped
    networks:
      - backend

Volumes and Networks

volumes:
  app-data:           # Named volume (Docker-managed)
  db-data:
    driver: local

networks:
  backend:
    driver: bridge
  frontend:
    driver: bridge
See: references/compose-patterns.md for full patterns including profiles, watch mode, and override files.

Security Quick Reference

AreaRecommendation
UserRun as non-root: RUN adduser -D appuser && USER appuser
Base imagePin digest: FROM python:3.12-slim@sha256:abc123...
FilesystemRead-only root: docker run --read-only --tmpfs /tmp
CapabilitiesDrop all, add needed: --cap-drop=ALL --cap-add=NET_BIND_SERVICE
SecretsBuildKit secrets: RUN --mount=type=secret,id=key cat /run/secrets/key
ScanningScan images: trivy image myapp:latest or grype myapp:latest
No latestAlways use specific tags and pin versions
Minimal imageUse distroless or scratch when possible
No SUIDRUN find / -perm /6000 -type f -exec chmod a-s {} +
NetworkUse internal networks for backend services

Non-Root User Pattern

# Debian/Ubuntu-based
RUN groupadd -r appuser && useradd -r -g appuser -d /app -s /sbin/nologin appuser
COPY --chown=appuser:appuser . /app
USER appuser

# Alpine-based
RUN addgroup -S appuser && adduser -S -G appuser appuser
COPY --chown=appuser:appuser . /app
USER appuser

# Distroless (built-in nonroot user)
FROM gcr.io/distroless/static:nonroot
USER nonroot:nonroot

Common Gotchas

GotchaProblemFix
Large imagesShipping build tools, node_modules in final imageMulti-stage builds
Cache bustingCOPY.. before RUN npm installCopy lockfile first, install, then copy source
Secrets in layersCOPY.env. or ARG SECRET=... bakes secrets into image historyUse --mount=type=secret or runtime env vars
PID 1 problemApp doesn't receive SIGTERM, zombie processesUse tini as init or exec form for CMD
TimezoneContainer uses UTCSet TZ env var or install tzdata
DNS cachingAlpine musl DNS issuesUse RUN apk add --no-cache libc6-compat or switch to slim
apt cacheapt-get update cached from old layerAlways combine update && install in one RUN
Missing signalsShell form (CMD npm start) wraps in /bin/shExec form: CMD ["node", "server.js"]
Build context sizeSending GB of data to daemonAdd .dockerignore, check with docker build --progress=plain
Layer explosionEach RUN creates a layerChain related commands with &&

PID 1 / Signal Handling Fix

# Option 1: Use tini as init process
RUN apt-get update && apt-get install -y --no-install-recommends tini \
    && rm -rf /var/lib/apt/lists/*
ENTRYPOINT ["tini", "--"]
CMD ["node", "server.js"]

# Option 2: Docker init flag (Docker 23.0+)
# docker run --init myapp

# Option 3: Node.js - handle signals in code
# process.on('SIGTERM', () => { server.close(); process.exit(0); });

Essential Docker Commands

# Build
docker build -t myapp:1.0 .
docker build -t myapp:1.0 --target production .    # Multi-stage target
docker build --no-cache -t myapp:1.0 .              # Force rebuild

# Run
docker run -d --name myapp -p 8080:8000 myapp:1.0
docker run --rm -it myapp:1.0 /bin/sh               # Interactive debug
docker run --read-only --tmpfs /tmp myapp:1.0        # Read-only root

# Inspect
docker images --format "table {{.Repository}}\t{{.Tag}}\t{{.Size}}"
docker history myapp:1.0                             # Layer breakdown
docker inspect myapp:1.0 | jq '.[0].Config'         # Image config

# Debug running container
docker exec -it myapp /bin/sh
docker logs -f myapp
docker stats myapp

# Cleanup
docker system prune -a --volumes                     # Remove everything unused
docker image prune -a                                # Remove unused images

Reference Files

FileContents
references/multi-stage-builds.mdPer-language multi-stage patterns (Go, Rust, Node, Python)
references/compose-patterns.mdCompose services, networking, profiles, watch, overrides
references/optimization.mdImage size, BuildKit, security scanning, debugging

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.8%
按下载量换算43

Claude

30.25%
按下载量换算36

Cursor

18.18%
按下载量换算21

Gemini CLI

8.42%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/0xdarkmatter/claude-mods --skill docker-ops 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills