Automate Skill - AIEmpire-Core

Purpose

Automate repetitive tasks within the AI Empire ecosystem. This skill handles creating, scheduling, and managing automated workflows across all system components.

Capabilities

1. GitHub Actions Automation

Create or modify workflows in .github/workflows/:

yaml

# Standard workflow template
name: [Workflow Name]
on:
  schedule:
    - cron: '[schedule]'
  workflow_dispatch:

jobs:
  run:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: [Step]
        run: |
          [commands]

Existing Workflows:

Workflow	Schedule	Purpose
`auto-content-generation.yml`	Every 4h	Generate X/Twitter content
`claude-health-check.yml`	Every 30min	Monitor Claude API status
`issue-command-bot.yml`	On issue comment	Process @bot commands
`revenue-tracking.yml`	Daily 9 AM	Revenue report
`x-auto-poster.yml`	Daily 7 AM	Post to X/Twitter
`weekly-review.yml`	Weekly	System review
`gold-nugget-extractor.yml`	On trigger	Extract insights
`daily-content-engine.yml`	Daily	Full content pipeline
`deploy-mobile.yml`	On push	Deploy mobile PWA

2. Atomic Reactor Tasks

Define new automated tasks in atomic-reactor/tasks/:

yaml

# Task template
name: [task-name]
description: [what it does]
schedule: [cron or trigger]
model: [ollama|kimi|claude]
priority: [1-5]
steps:
  - action: [action-type]
    params:
      key: value

3. OpenClaw Cron Jobs

Modify or add cron jobs via openclaw-config/jobs.json:

json

{
  "name": "[job-name]",
  "schedule": "*/30 * * * *",
  "skill": "[skill-name]",
  "model": "kimi-k2.5",
  "enabled": true
}

4. Workflow System Automation

Leverage the 5-step compound loop:

bash

# Automate a full cycle
python workflow-system/orchestrator.py

# Schedule the cowork daemon
python workflow-system/cowork.py --daemon --interval 1800

# Focus on specific area
python workflow-system/cowork.py --daemon --focus revenue

5. N8N Workflows

Connect external services via n8n-workflows/n8n_connector.py.

Automation Patterns

Pattern: Content Pipeline

code

Trigger: Cron (daily 7 AM)
Steps:
  1. Trend scan (Ollama) → trending topics
  2. Content generation (Kimi) → 10 posts
  3. Quality filter (Ollama) → top 5
  4. Schedule posting (GitHub Action)
  5. Track engagement (CRM)

Pattern: Lead Qualification

code

Trigger: New lead in CRM
Steps:
  1. Enrich lead data (web scrape)
  2. BANT scoring (Ollama)
  3. If score > 7: route to Sales Agent
  4. If score 4-7: add to nurture sequence
  5. If score < 4: add to content funnel

Pattern: Revenue Tracking

code

Trigger: Cron (daily 9 AM)
Steps:
  1. Pull Gumroad sales
  2. Pull Fiverr earnings
  3. Aggregate totals
  4. Compare to targets
  5. Generate report → docs/kpi/
  6. Alert if below target

Cost Rules

Model	Use For	Cost
Ollama (local)	95% of tasks - routine, bulk, filtering	FREE
Kimi K2.5	Volume content, swarm tasks	~0.001 EUR/call
Claude Haiku	Quick quality checks	~0.01 EUR/call
Claude Sonnet	Code review, strategy	~0.05 EUR/call
Claude Opus	Critical decisions only	~0.15 EUR/call

Budget limit: 100 EUR/month total. Always default to Ollama first.

Resource Guard Integration

Before creating compute-heavy automations, check thresholds:

Level	CPU	RAM	Action
NORMAL	< 70%	< 75%	Full concurrency
WARN	> 70%	> 75%	Reduce to 200 concurrent
CRITICAL	> 85%	> 85%	Reduce to 50, outsource to Kimi
EMERGENCY	> 95%	> 92%	Pause all agents

Creating New Automations

•Define the trigger (cron, webhook, event, manual)
•Choose the execution layer (GitHub Actions, Atomic Reactor, OpenClaw, n8n)
•Select models by cost (Ollama first)
•Set resource limits
•Add error handling and fallback
•Test with --test flag before full deployment
•Monitor via KPI snapshots in docs/kpi/