The Kiro Automation Stack — Interactive Explainer

🏗️ Why a Stack?

Building an AI-powered workflow isn't one thing — it's layers. Each layer solves a different problem. You can use one layer alone, or combine all five for a fully autonomous agent.

📏

Steering

Global rules that apply to every AI interaction. Like a company policy document.

🧠

Skills

On-demand expertise that activates when needed. Like a specialist you call in.

⚡

Hooks

Auto-triggers that fire on events. Like a motion sensor that turns on the lights.

🤖

Agents

Autonomous colleagues you delegate to. Like assigning a task to a team member.

💡

The key insight: Steering sets the rules. Skills define the expertise. Hooks trigger automatically. Agents work independently. MCP connects to your data. Together = an autonomous workflow.

🎯 The Stack at a Glance

📏

Steering

Always-on rules

🧠

Skills

On-demand templates

⚡

Hooks

Event triggers

🤖

Agents

Delegated workers

↕️

🔌

MCP — Model Context Protocol

Connects AI to databases, APIs, and files

🔍 Explore Each Layer

Click a layer to see details, file location, when to use it, and a real example.

📏

Steering

🧠

Skills

⚡

Hooks

🤖

Agents

🔌

MCP

📏 Steering — Global Rules

.kiro/steering/*.md

Steering files are always-on instructions loaded into every conversation. They're like a company policy that every AI interaction must follow — regardless of what task is being performed.

When to use:

Rules that apply to ALL tasks (currency, PII, company name)
Formatting standards your team always follows
Compliance constraints that never change
Context that every conversation needs (market, audience)

When NOT to use:

Task-specific instructions (use Skills instead)
Large reference documents (use manual-inclusion steering)
Anything that only applies sometimes

Example:

---
inclusion: always
---
# AnyCompany Rules
1. All amounts in SGD unless specified
2. Never include PII — use [REDACTED]
3. Risk ratings: GREEN/AMBER/RED only
4. Cite specific data for every claim

🧠 vs 🔌 Skills vs MCP — When to Use Which?

This is the most common question on Day 3. Skills and MCP both make AI more capable — but they do completely different things. Here's the clearest way to think about it:

🧠

Skills = What to say

Skills tell the AI how to think and what to produce. They're prompt templates with personas, output formats, guardrails, and decision rules.

Analogy: A recipe that tells the chef what dish to make, what ingredients to use, and how to plate it.

🔌

MCP = Where to get data

MCP gives the AI access to external data and systems. It connects to databases, APIs, and files so the AI can query real data instead of relying on what you paste in.

Analogy: The pantry and fridge that the chef pulls ingredients from. Without it, the chef can only cook with what you hand them.

🏦 AnyCompany Example: Merchant Risk Assessment

Watch how Skills and MCP play different roles in the same workflow:

Step	What happens	Powered by	Why
1. Get data	Query merchant database for transaction history, chargeback rates, complaints	🔌 MCP	AI needs access to live data it can't see otherwise
2. Analyze	Apply the 8-section risk assessment template with GREEN/AMBER/RED ratings	🧠 Skill	AI needs instructions on what format and rules to follow
3. Cross-check	Query compliance_alerts table for open alerts on this merchant	🔌 MCP	AI needs more data from a different table
4. Decide	Apply decision rules: chargeback >2% = RED, complaints >5 = escalate	🧠 Skill	AI needs business rules to make the right call
5. Report	Generate the structured assessment with citations	🧠 Skill	AI needs the output template to format correctly

💡

The pattern: MCP fetches the data. Skills process the data. They alternate throughout a workflow — MCP brings ingredients, Skills cook the meal.

❓ Quick Decision: Skill or MCP?

If you need the AI to...	Use	Example
Follow a specific output format	🧠 Skill	8-section risk report with GREEN/AMBER/RED
Query a database	🔌 MCP	Pull merchant transaction history
Use a specific persona	🧠 Skill	"You are a Senior Risk Analyst..."
Read files from a folder	🔌 MCP	Scan invoices/ directory for new PDFs
Apply business rules and guardrails	🧠 Skill	Chargeback >2% = RED, escalate if >$25K
Send a Slack message or call an API	🔌 MCP	Post alert to #risk-alerts channel
Generate a structured report	🧠 Skill	Credit committee narrative with bull/bear case
Access live data it can't see	🔌 MCP	Check compliance status in real-time

🔑

One-liner: Skills = the brain (how to think). MCP = the hands (how to reach data). You almost always need both for a production workflow. A skill without MCP can only work with data you paste in. MCP without a skill gives raw data with no structure.

🧭 Which Layer Do I Need?

Answer the questions below to find the right layer for your use case.

Does this rule/behavior apply to EVERY conversation, or only specific tasks?

Every conversation

It's a global standard

Specific tasks only

Only when doing certain work

🔄

Try different paths! Click through the decision tree multiple times to see how different needs map to different layers. The tree resets when you click "Start over".

🎬 Real-World Scenarios

See how the layers combine for common AnyCompany Finance workflows. Click a scenario to see the breakdown.

A new merchant application arrives. The system automatically assesses risk from multiple perspectives and produces a committee-ready brief.

📊 Side-by-Side Comparison

Aspect	Steering	Skills	Hooks	Agents	MCP
Scope	Global — all tasks	On-demand — specific tasks	Event-driven — automatic	Delegated — independent	Data layer — connections
Activation	Always loaded	Auto-activates by keyword	Fires on IDE events	Invoked by name	Available when configured
File location	`.kiro/steering/`	`.kiro/skills/*/`	`.kiro/hooks/`	`.kiro/agents/`	`.kiro/settings/mcp.json`
Token cost	Every message	Only when active	Zero (triggers only)	Own context window	Per tool call
Analogy	Company policy	Specialist consultant	Motion sensor	Team member	Database connection
Best for	Standards, constraints	Structured outputs	Automation triggers	Review, validation	Data access
AnyCompany example	"All amounts in SGD"	Risk assessment template	"Run assessment on new CSV"	Finance reviewer agent	Merchant database

💡

Key distinction — Skills vs Agents: A skill is a recipe Kiro follows (instructions added to its context). An agent is a separate worker Kiro delegates to (runs independently with its own tools). Use skills for structured output generation. Use agents for review, validation, or tasks needing independent judgment.

🔗 The Full Pipeline

Here's how all 5 layers work together in a complete autonomous workflow:

1
Steering loads global rules
Every conversation starts with your team's standards: SGD currency, no PII, GREEN/AMBER/RED ratings, cite all data.
Steering

2

Hook detects a trigger event

A new CSV file appears in the workspace. The hook fires and sends a message to Kiro: "New merchant data — run risk assessment."

Hook

3

Skill activates with the template

Kiro recognizes "risk assessment" and loads the merchant-risk-assessment SKILL.md — the 8-section template with decision rules and guardrails.

Skill

4

MCP connects to data sources

The skill needs merchant transaction history. MCP queries the SQLite database and returns 6 months of data — no manual copy-paste needed.

MCP

5

Agent reviews the output

Kiro delegates to the finance-reviewer agent. The agent independently checks completeness, guardrail compliance, and scores quality 1-10.

Agent

✓
Result: Committee-ready assessment
A complete, reviewed, quality-scored risk assessment — generated automatically from a single CSV file drop. No prompting required.

🎯

You don't need all 5 layers. Start with Steering alone (Day 2 level). Add Skills when you want reusable templates. Add Hooks when you want automation. Add MCP when you need data access. Add Agents when you need independent review. Build up gradually.