Token导航 LogoToken导航TokenDH.com
研究检索敏感数据github未标认证来源可访问许可证需确认审计通过

ansible-idempotencyansible 幂等性

Agent Skill

ansible-idempotency 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

408

周安装

17

GitHub Stars

18

下载量

136
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:ansible-idempotency(ansible 幂等性)
来源仓库:https://github.com/basher83/lunar-claude
仓库路径:skills/ansible-idempotency
安装命令:
npx skills add https://github.com/basher83/lunar-claude --skill ansible-idempotency
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/basher83/lunar-claude --skill ansible-idempotency

简介

ansible-idempotency 专注于确保 Ansible 任务真正幂等运行。

  • 提供 changed_when 和 failed_when 等关键控制手段。
  • 适用于需要重复执行而不产生副作用的自动化场景。
  • 安装前需确认权限范围、维护状态及是否涉及命令执行或状态判断。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Ansible Idempotency Patterns

Techniques for ensuring Ansible tasks are truly idempotent - producing the same result whether run once or multiple times.

Core Directives

changed_when

Controls when Ansible reports a task as "changed". Critical for command and shell modules which always report changed by default.

- name: Check if service exists
  ansible.builtin.command: systemctl status myservice
  register: service_check
  changed_when: false  # Read-only operation, never changes anything

failed_when

Controls when Ansible considers a task failed. Allows graceful handling of expected errors.

- name: Check resource existence
  ansible.builtin.command: check-resource {{ resource_id }}
  register: check_result
  failed_when: false  # Don't fail, we'll check the result ourselves

register

Captures task output for use in changed_when and failed_when expressions.

- name: Run command
  ansible.builtin.command: some-command
  register: cmd_result
  # Now cmd_result.rc, cmd_result.stdout, cmd_result.stderr are available

Pattern 1: Detect Actual Changes

Make commands report "changed" only when something actually changed:

- name: Create Proxmox API token
  ansible.builtin.command: >
    pveum user token add {{ username }}@pam {{ token_name }}
  register: token_result
  changed_when: "'already exists' not in token_result.stderr"
  failed_when:
    - token_result.rc != 0
    - "'already exists' not in token_result.stderr"
  no_log: true

Key pattern: Detect specific output that indicates no change occurred.

Pattern 2: Check Before Create

Check if a resource exists before creating it:

- name: Check if VM template exists
  ansible.builtin.shell: |
    set -o pipefail
    qm list | awk '{print $1}' | grep -q "^{{ template_id }}$"
  args:
    executable: /bin/bash
  register: template_exists
  changed_when: false  # Checking doesn't change anything
  failed_when: false   # Not finding it isn't a failure

- name: Create VM template
  ansible.builtin.command: >
    qm create {{ template_id }}
    --name {{ template_name }}
    --memory 2048
  when: template_exists.rc != 0  # Only create if doesn't exist
  register: create_result
  changed_when: create_result.rc == 0

Pattern 3: Verify After Create

Confirm resource creation succeeded:

- name: Create VM
  ansible.builtin.command: >
    qm create {{ vmid }} --name {{ vm_name }}
  register: create_result
  changed_when: true

- name: Verify VM was created
  ansible.builtin.shell: |
    set -o pipefail
    qm list | grep "{{ vmid }}"
  args:
    executable: /bin/bash
  register: verify_result
  changed_when: false
  failed_when: verify_result.rc != 0

Pattern 4: Conditional Change Detection

Use output content to determine if change occurred:

- name: Update cluster configuration
  ansible.builtin.command: update-config --apply
  register: update_result
  changed_when: "'Configuration updated' in update_result.stdout"
  failed_when: "'Error' in update_result.stderr"

Common Patterns

Output Indicatorchanged_when Expression
"already exists""'already exists' not in result.stderr"
"no changes""'no changes' not in result.stdout"
"created""'created' in result.stdout"
"updated""'updated' in result.stdout"
Exit code 0 = createdresult.rc == 0

Pattern 5: Multiple Failure Conditions

Allow specific "failures" that are actually expected:

- name: Run database migration
  ansible.builtin.command: /usr/bin/migrate-database
  register: migrate_result
  failed_when:
    - migrate_result.rc != 0
    - "'already applied' not in migrate_result.stdout"
    - "'no pending migrations' not in migrate_result.stdout"
  changed_when: "'applied' in migrate_result.stdout and 'already' not in migrate_result.stdout"

Pattern 6: Read-Only Operations

Mark read-only operations as never changed:

# Checking status
- name: Get cluster status
  ansible.builtin.command: pvecm status
  register: cluster_status
  changed_when: false
  failed_when: false

# Gathering information
- name: List available images
  ansible.builtin.command: qm list
  register: vm_list
  changed_when: false

# Verification checks
- name: Verify service is running
  ansible.builtin.command: systemctl is-active nginx
  register: nginx_status
  changed_when: false
  failed_when: false

Pattern 7: Retry Until Success

Use until for operations that may need retries:

- name: Wait for service to be ready
  ansible.builtin.uri:
    url: http://localhost:8080/health
    status_code: 200
  register: health_check
  until: health_check.status == 200
  retries: 30
  delay: 10
  # Total wait: up to 5 minutes

With command:

- name: Wait for VM to get IP address
  ansible.builtin.command: qm agent {{ vmid }} network-get-interfaces
  register: vm_network
  until: vm_network.rc == 0
  retries: 12
  delay: 5
  changed_when: false

Pattern 8: Set Facts for State

Use facts to track state across tasks:

- name: Check existing cluster status
  ansible.builtin.command: pvecm status
  register: cluster_status
  failed_when: false
  changed_when: false

- name: Set cluster facts
  ansible.builtin.set_fact:
    is_cluster_member: "{{ cluster_status.rc == 0 }}"
    in_target_cluster: "{{ cluster_name in cluster_status.stdout }}"

- name: Create cluster
  ansible.builtin.command: pvecm create {{ cluster_name }}
  when: not in_target_cluster
  register: cluster_create
  changed_when: cluster_create.rc == 0

Anti-Patterns to Avoid

Always Changed

# BAD - Always shows changed
- name: Check status
  ansible.builtin.command: systemctl status app

# GOOD
- name: Check status
  ansible.builtin.command: systemctl status app
  register: status_check
  changed_when: false
  failed_when: false

Silent Failure Suppression

# BAD - Hides all errors
- name: Critical operation
  ansible.builtin.command: important-command
  failed_when: false

# GOOD - Only allow expected "errors"
- name: Critical operation
  ansible.builtin.command: important-command
  register: result
  failed_when:
    - result.rc != 0
    - "'expected condition' not in result.stderr"

No Output Capture

# BAD - Can't check results
- name: Run command
  ansible.builtin.command: create-resource

# GOOD
- name: Run command
  ansible.builtin.command: create-resource
  register: result
  changed_when: "'created' in result.stdout"

Shell Script Requirements

Use strict error handling in shell scripts:

- name: Run pipeline
  ansible.builtin.shell: |
    set -euo pipefail
    cat data.txt | grep pattern | sort | uniq
  args:
    executable: /bin/bash
  register: pipeline_result
  changed_when: false

Why set -euo pipefail?

FlagPurpose
-eExit on any command failure
-uError on undefined variables
-o pipefailCatch errors in pipelines

Testing Idempotency

Verify playbooks are idempotent by running twice:

# First run - may show changes
uv run ansible-playbook playbooks/setup.yml

# Second run - should show 0 changes
uv run ansible-playbook playbooks/setup.yml

# If second run shows changes, playbook is NOT idempotent

Common changed_when Expressions

# Never changed (read-only)
changed_when: false

# Always changed (one-time operations)
changed_when: true

# Based on output content
changed_when: "'created' in result.stdout"
changed_when: "'already exists' not in result.stderr"
changed_when: "'updated' in result.stdout"

# Based on return code
changed_when: result.rc == 0
changed_when: result.rc != 1

# Complex conditions
changed_when:
  - result.rc == 0
  - "'no changes' not in result.stdout"

Utility Script

Use the idempotency checker to analyze playbooks for common issues:

# Check a single playbook
${CLAUDE_PLUGIN_ROOT}/skills/ansible-idempotency/scripts/check_idempotency.py ansible/playbooks/my-playbook.yml

# Check multiple playbooks
${CLAUDE_PLUGIN_ROOT}/skills/ansible-idempotency/scripts/check_idempotency.py ansible/playbooks/*.yml

# Strict mode (info issues become warnings)
${CLAUDE_PLUGIN_ROOT}/skills/ansible-idempotency/scripts/check_idempotency.py --strict ansible/playbooks/my-playbook.yml

# Summary only
${CLAUDE_PLUGIN_ROOT}/skills/ansible-idempotency/scripts/check_idempotency.py --summary ansible/playbooks/*.yml

The script detects:

  • Command/shell tasks without changed_when
  • Shell tasks without set -euo pipefail
  • Tasks missing no_log that may contain secrets
  • Tasks missing name attribute
  • Use of deprecated short module names (non-FQCN)

Script location: ${CLAUDE_PLUGIN_ROOT}/skills/ansible-idempotency/scripts/check_idempotency.py

Related Skills

  • ansible-error-handling - Block/rescue patterns
  • ansible-fundamentals - Module selection (prefer native modules)
  • ansible-proxmox - Proxmox-specific idempotency patterns

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

36.81%
按下载量换算50

Claude

28.6%
按下载量换算39

Cursor

21.81%
按下载量换算30

Gemini CLI

9.08%
按下载量换算12

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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