Alex LaGuardia

Full-stack software engineer. Building AI systems, game engines, and production platforms. Python, Rust, TypeScript. Currently shipping a cognitive AI layer and a multi-strategy trading system.

Joined on Mar 18, 2026

Alex LaGuardia

Jul 18

An AI agent exported a patient record. Your logs can't say who told it to.

#mcp #security #ai #agents

6 min read

Want to connect with Alex LaGuardia?

Create an account to connect with Alex LaGuardia. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

Alex LaGuardia

Jul 17

The model refused. Invisible characters changed its mind.

#mcp #security #ai #agents

5 min read

Alex LaGuardia

Jul 12

A shipped agent framework ran the hidden payload. My own scanner caught it.

#mcp #security #ai #agents

5 min read

Alex LaGuardia

Jul 10

The tool swore it never exports data. The model read a different field.

#mcp #security #ai #agents

4 min read

Alex LaGuardia

Jul 8

ChatGPT asks permission before it acts. It just doesn't count a GET as acting.

#mcp #security #ai #agents

5 min read

Alex LaGuardia

Jul 6

Scan the tool description, I said. So I hid the payload somewhere else.

#mcp #security #ai #agents

4 min read

Alex LaGuardia

Jul 6

I put the same attack in two places. The one everyone warns about was the safe one.

#mcp #security #ai #agents

4 min read

Alex LaGuardia

Jul 4

I pinned each issuer's public key. Then the IdP rotated it.

#ai #security #oauth #mcp

7 min read

Alex LaGuardia

Jul 3

An AI agent acted across two companies. Whose audit log knows which human?

#ai #security #oauth #mcp

6 min read

Alex LaGuardia

Jun 27

I broke my own governed MCP server by hand, then built the scanner that catches the class

#ai #security #mcp #opensource

5 min read

Alex LaGuardia

Jun 24

An AI agent acted across two companies. Whose audit log knows which human?

#ai #security #oauth #mcp

6 min read

Alex LaGuardia

Jun 23

An AI agent exported a patient record. Your logs can't say who told it to.

#ai #security #llm #mcp

5 min read

Alex LaGuardia

Jun 10

Same question, three answers: a governed MCP server with receipts

#ai #mcp #python #llm

3 min read

Alex LaGuardia

May 1

Your AI Agent Forgets Everything Between Sessions (Here's How to Fix It)

#ai #python #claudecode #productivity

3 min read

Alex LaGuardia

Apr 30

Your MCP Servers Are Flying Blind (Here's How to Fix It)

#mcp #ai #python #monitoring

2 min read

Alex LaGuardia

Apr 4

Three freelancer tools died in two months. I built the replacement.

#python #fastapi #ai #freelancing

3 min read

Alex LaGuardia

Mar 31

I Built a Security Scanner That Uses AI to Review Its Own Findings

#security #ai #python #opensource

4 min read

Alex LaGuardia

Mar 26

OAuth2, Two APIs, and Soft Deletes — Building an MCP Server for FreshBooks

#mcp #ai #python #freshbooks

5 min read

Alex LaGuardia

Mar 26

Zero to 33 Tools: Building the First MCP Server for ActiveCampaign

#mcp #ai #python #activecampaign

4 min read

Alex LaGuardia

Mar 25

How I Built a Full Product in One Night with 3 Parallel AI Agents

#ai #productivity #python #opensource

3 min read

Alex LaGuardia

Mar 25

I Built a Nervous System for AI Agents (Not Another Memory Store)

#ai #python #mcp #opensource

5 min read

Alex LaGuardia

Mar 25

Managing a WooCommerce Store from Claude — 34 MCP Tools I Wish Existed

#mcp #ai #python #woocommerce

4 min read

Alex LaGuardia

Mar 24

33 Tools for Mailchimp in One MCP Server — Here's How I Built It

#mcp #ai #python #mailchimp

4 min read

Alex LaGuardia

Mar 19

I built a CLI inspector for MCP servers

#mcp #python #cli #opensource

1 min read

Alex LaGuardia

Mar 18

How I Built a Cognitive AI Layer That Routes Thoughts to the Right Brain

#ai #python #architecture #llm

5 min read

DEV Community

Alex LaGuardia

Badges

Writing Debut

An AI agent exported a patient record. Your logs can't say who told it to.

Want to connect with Alex LaGuardia?

The model refused. Invisible characters changed its mind.

A shipped agent framework ran the hidden payload. My own scanner caught it.

The tool swore it never exports data. The model read a different field.

ChatGPT asks permission before it acts. It just doesn't count a GET as acting.

Scan the tool description, I said. So I hid the payload somewhere else.

I put the same attack in two places. The one everyone warns about was the safe one.

I pinned each issuer's public key. Then the IdP rotated it.

An AI agent acted across two companies. Whose audit log knows which human?

I broke my own governed MCP server by hand, then built the scanner that catches the class

An AI agent acted across two companies. Whose audit log knows which human?

An AI agent exported a patient record. Your logs can't say who told it to.

Same question, three answers: a governed MCP server with receipts

Your AI Agent Forgets Everything Between Sessions (Here's How to Fix It)

Your MCP Servers Are Flying Blind (Here's How to Fix It)

Three freelancer tools died in two months. I built the replacement.

I Built a Security Scanner That Uses AI to Review Its Own Findings

OAuth2, Two APIs, and Soft Deletes — Building an MCP Server for FreshBooks

Zero to 33 Tools: Building the First MCP Server for ActiveCampaign

How I Built a Full Product in One Night with 3 Parallel AI Agents

I Built a Nervous System for AI Agents (Not Another Memory Store)

Managing a WooCommerce Store from Claude — 34 MCP Tools I Wish Existed

33 Tools for Mailchimp in One MCP Server — Here's How I Built It

I built a CLI inspector for MCP servers

How I Built a Cognitive AI Layer That Routes Thoughts to the Right Brain