auto-flow
Skill1 starsChanged 3 months agoMITView on GitHub
Service deployment and workflow orchestration skill for FlowMind. Deploy/release services (部署/发布/上线) via Feiliu pipeline MCP, query pipeline status and runs (流水线/飞流), and orchestrate multi-step workflows.
In this repository30 other files it gives its agents.
- api-syncSkill
- archive-changeSkill
- build-reportSkill
- change-impact-reviewSkill
- code-review-auditSkill
- code-reviewSkill
- coverage-reportSkill
- data-logic-validationSkill
- data-risk-auditSkill
- data-validationSkill
- dingtalk-notifySkill
- git-reviewSkill
- knowledge-opsSkill
- learning-engineSkill
- learning-feedbackSkill
- log-auditSkill
- offline-design-analysisSkill
- project-reviewSkill
- release-inspectionSkill
- report-insightsSkill
- requirement-analystSkill
- resource-bindSkill
- service-governance-auditSkill
- service-governance-changeSkill
- service-governance-diagnosisSkill
- service-management-log-analysisSkill
- sls-log-auditSkill
- test-reportSkill
- yapi-sync-interfaceSkill
- yuque-sync-designSkill
The file
Files
Auto Flow Skill
Deploy services via Feiliu (飞流) CI/CD pipelines over MCP, query pipeline status, and orchestrate multi-step workflows.
Features
🔄 Workflow Execution
- Sequential steps
- Parallel execution
- Conditional branching
- Error handling
📋 Workflow Templates
- Reusable workflows
- Parameterized steps
- Version control
- Team sharing
📊 Monitoring
- Step tracking
- Progress reporting
- Error logging
- Performance metrics
Trigger Patterns
"自动化", "automation", "workflow"
"执行流程", "运行流程", "启动流程", "pipeline"
"批量", "batch"
"定时", "scheduled"
"工作流", "work flow"
"部署", "deploy", "deployment"
"发布", "上线"
"流水线", "飞流"Workflow Definition
YAML Format
name: deploy-pipeline
description: Deploy to production
steps:
- name: test
action: run-tests
params:
coverage: true
- name: build
action: build-artifact
depends_on: [test]
- name: deploy-staging
action: deploy
params:
environment: staging
depends_on: [build]
- name: integration-test
action: run-integration-tests
depends_on: [deploy-staging]
- name: deploy-prod
action: deploy
params:
environment: production
depends_on: [integration-test]
when: "{{branch}} == 'main'"Output Format
┌─────────────────────────────────────────────────────┐
│ Workflow Execution │
├─────────────────────────────────────────────────────┤
│ Workflow: {name} │
│ Status: {status} │
│ Progress: {progress}% │
├─────────────────────────────────────────────────────┤
│ Steps: │
│ ✓ test (2m 30s) │
│ ✓ build (1m 15s) │
│ ✓ deploy-staging (45s) │
│ ⏳ integration-test (running...) │
│ ○ deploy-prod (pending) │
├─────────────────────────────────────────────────────┤
│ Duration: 4m 30s │
│ ETA: 2m 15s │
└─────────────────────────────────────────────────────┘Built-in Workflows
Development Workflow
name: dev-workflow
steps:
- name: code-review
skill: code-review
- name: test
skill: data-validation
- name: docs
skill: api-sync
- name: archive
skill: archive-changeDeployment Workflow
name: deploy-workflow
steps:
- name: validate
action: validate-config
- name: backup
action: backup-database
- name: deploy
action: deploy-application
- name: verify
action: health-check
- name: notify
action: send-notificationParallel Execution(全局唯一 → 流水线各自并行 → 结果合并唯一)
部署清单解析出 ≥2 个服务,或用户一次要求部署多个服务时:
- 全局唯一:一次部署只有一份部署报告与一个总体状态,由主 agent 汇总;环境以用户指定为准,禁止混环境。
- 流水线各自并行:逐服务解析 pipelineId 并保留
{env}-{serviceName}前缀校验(不匹配即拒绝该服务并说明原因);优先flowStartBatchPipelineRun批量启动,或并发调用flowStartPipelineRun(≤4 路);状态按 runId 并行轮询互不阻塞。 - 结果合并唯一:全部到终态后汇总各服务三阶段耗时与产物镜像 → 唯一部署报告(成功/部分成功/失败 + 失败阶段与日志入口)→ 执行记录统一写一次。
- 失败隔离:单服务失败不阻塞其他服务;同服务禁止并行两条流水线,只能串行重试(
flowRetryPipelineJobRun)。
宿主支持并行子 agent 时按服务并行派发与轮询;flowmind 运行时内以 flowStartBatchPipelineRun + 并发状态查询实现,输出结构不变。
Learning Integration
This skill supports FlowMind learning:
- Workflow Preferences: Learns common workflows
- Step Order: Learns preferred step sequence
- Error Handling: Learns recovery preferences
User: "部署前先备份数据库"
FlowMind: ✓ Learned: Backup before deploy
User: [Next deploy]
FlowMind: [Includes backup step automatically]Examples
Example 1: Execute Workflow
User: 执行部署流程
FlowMind:
┌─────────────────────────────────────────────────────┐
│ Workflow: deploy-production │
├─────────────────────────────────────────────────────┤
│ ✓ validate (5s) │
│ ✓ backup (2m 15s) │
│ ✓ deploy (3m 30s) │
│ ✓ verify (30s) │
│ ✓ notify (5s) │
├─────────────────────────────────────────────────────┤
│ Status: ✓ Completed │
│ Duration: 6m 25s │
│ Environment: production │
└─────────────────────────────────────────────────────┘Example 2: Custom Workflow
User: 创建一个代码提交流程
FlowMind:
┌─────────────────────────────────────────────────────┐
│ Workflow Created: code-commit │
├─────────────────────────────────────────────────────┤
│ Steps: │
│ 1. lint - Run linter │
│ 2. test - Run tests │
│ 3. review - Code review │
│ 4. commit - Create commit │
│ 5. push - Push to remote │
├─────────────────────────────────────────────────────┤
│ Usage: flowmind run code-commit │
└─────────────────────────────────────────────────────┘Configuration
{
"auto-flow": {
"workflows": {
"path": "./workflows",
"autoDiscover": true
},
"execution": {
"parallel": true,
"maxConcurrent": 4,
"timeout": 3600
},
"notifications": {
"onComplete": true,
"onError": true
}
}
}Workflow Actions
Built-in Actions
| Action | Description |
run-command | Execute shell command |
run-tests | Run test suite |
deploy | Deploy application |
notify | Send notification |
wait | Wait for condition |
approve | Request approval |
Custom Actions
{
"auto-flow": {
"actions": {
"my-action": {
"command": "npm run my-script",
"timeout": 300
}
}
}
}Error Handling
Retry Configuration
steps:
- name: deploy
action: deploy
retry:
max: 3
delay: 5000
backoff: exponentialFailure Actions
steps:
- name: deploy
action: deploy
on_failure:
- action: rollback
- action: notify
params:
message: "Deploy failed!"Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

