Back to Documentation

Real Examples

Production examples from running 11 squads, 69 agents, and $758/week in AI costs.

These are not hypothetical examples. Every output below comes from a production system that runs Agents Squads using squads-cli.

11
Active Squads
69
Total Agents
511
Commits (30d)
$758
AI Costs/Week

Engineering Squad

6 agents · 224 commits

Automates pull request reviews, runs test suites, and maintains deployment pipelines.

Use Case
CI/CD automation and code review
Key Workflows
PR Review
Automated code quality checks on every pull request
Test Runner
Executes test suites and reports failures
Deployment
Manages staging and production deployments
Real production output 
$ squads run engineering

✓ Reviewed 3 PRs for code quality
✓ Ran test suite: 147/147 passing
✓ Updated deployment docs
✓ Synced CI/CD config across repos

Intelligence Squad

24 agents · 13 commits

Monitors enterprise data platforms, tracks competitor moves, and identifies market opportunities.

Use Case
Market research and competitive analysis
Key Workflows
Market Scanner
Daily monitoring of competitor pricing and features
Adoption Tracker
Tracks enterprise adoption patterns and customer counts
Signal Hunter
Identifies buying signals from job postings and tech blogs
Real production output 
$ squads memory query "enterprise adoption"

5 results found

◇ Snowflake Q4 2024: 9,437 customers (+33% YoY)
  └ intelligence/snowflake-monitor
◇ Databricks funding: $10B at $62B valuation
  └ intelligence/databricks-monitor
◇ dbt Labs enterprise tier pricing increased 40%
  └ intelligence/dbt-monitor

Customer Squad

4 agents · 4 commits

Identifies qualified leads, scores them based on fit, and generates personalized outreach.

Use Case
Lead qualification and outreach automation
Key Workflows
Lead Discovery
Finds companies matching ICP criteria
Lead Scoring
Scores leads based on fit, urgency, and budget signals
Outreach Gen
Creates personalized outreach based on company context
Real production output 
$ squads run customer/lead-scorer

Scored 47 leads from last week:

HOT (5):
  • Cursor (AI IDE) - enterprise tier, 50k+ users
  • Sierra (AI customer service) - raised $110M
  • Distyl (AI research) - targeting F500

WARM (12):
  • 8 companies with recent AI tooling job posts
  • 4 companies evaluating agentic frameworks

$ squads run customer/outreach-generator

Generated personalized emails for 5 HOT leads
Drafted LinkedIn connection messages
Created follow-up sequences

Website Squad

10 agents · 222 commits

Maintains website content, optimizes for search engines, and tracks analytics.

Use Case
Content updates and SEO optimization
Key Workflows
SEO Audit
Weekly SEO analysis and recommendations
Content Updates
Keeps documentation in sync with product changes
Analytics Review
Tracks visitor behavior and conversion metrics
Real production output 
$ squads run website/seo-optimizer

Analyzed 23 pages for SEO:

Recommendations:
  • /cli page: Add meta description
  • /docs/* pages: Missing H1 tags (3 pages)
  • Blog posts: Update internal linking

$ squads run website/content-updater

Updated pages:
  ✓ /cli.astro - Added real production data
  ✓ /docs/installation.astro - New step-by-step guide
  ✓ /docs/commands.astro - Full command reference

Git activity: 222 commits in 30 days

Marketing Squad

4 agents · 222 commits

Owns the marketing funnel from awareness to conversion.

Use Case
Content marketing and funnel optimization
Key Workflows
Content Calendar
Plans and schedules blog posts and social media
SEO Tracking
Monitors keyword rankings and organic traffic
Funnel Analytics
Tracks metrics at each funnel stage
Real production output 
$ squads goal list

marketing
Own the funnel: Awareness → Consideration → Conversion

● [8] Rank #1 for 'agents squads' on Google
● [9] Publish 4 SEO-optimized blog posts per month
● [10] Get 10 backlinks from AI/tech blogs

Progress:
  • Google ranking: #3 → #2 (this week)
  • Blog posts: 2/4 published this month
  • Backlinks: 3/10 acquired

Research Squad

6 agents · 14 commits

Transforms intelligence into insights about autonomous systems and agent orchestration.

Use Case
Autonomous AI systems research
Key Workflows
Paper Review
Monitors arXiv and research publications
Insight Synthesis
Converts intelligence into actionable insights
Pattern Analysis
Identifies emerging patterns in autonomous systems
Real production output 
$ squads memory show research

Recent insights:

◆ Anthropic Claude Code 80.9% SWE-bench score
  └ Implications for autonomous development workflows

◆ Multi-agent frameworks comparison
  └ AutoGen vs LangGraph vs custom orchestration

◆ Production agent patterns
  └ When to use workflows vs agents (Anthropic guidelines)

Next: Publish first research paper on agent orchestration patterns

How It Works

1

Squads = Domain Teams

Each squad owns a domain (engineering, intelligence, customer, etc.) and contains specialized agents.

2

Agents = Markdown Files

Agents are simple .md files with prompts and instructions. No microservices, no complex infrastructure.

3

Memory = Cross-Session State

Agents remember context across sessions. squads memory query "analytics" finds relevant past work.

4

Triggers = Autonomous Execution

Agents can run automatically based on Postgres conditions (cost alerts, metrics thresholds, etc.).

5

GitHub-Native Orchestration

Uses GitHub Issues for tasks, PRs for outputs, and Actions for automation. No custom orchestration layer.

Try it yourself

Install squads-cli and run your first squad in under 5 minutes.

Installation Guide Command Reference
Quick start:
$ npm install -g squads-cli
$ squads init
$ squads run engineering