16 de nov. de 2026
Build and test an MCP capacity calculator: missing inputs, explicit assumptions and AWS hosting
Run a local MCP capacity calculator, inspect how it preserves missing inputs and labels assumptions, and test real protocol responses over stdio and HTTP. Includes code, a reproducible lab and the existing AWS hosting boundary.
03 de nov. de 2026
AWS Builder Loft São Paulo: a new home for building together
My reflections from the October 29 opening: why a permanent home for builders matters for Brazil, and how one conversation can become shared learning, a practical experiment or a new collaboration.
27 de out. de 2026
Your agent crashed after the write. Should you retry?
A reproducible SQLite lab: crash after a write, reconcile using a receipt, and block retries when the evidence is insufficient.
27 de out. de 2026
Give your AWS architecture agent tools you can test: an open-source MCP workbench
Try 48 architecture tools for capacity, queues, IaC reviews and agent safety. Two reliability fixes show why deterministic tools still need explicit limits and regression tests.
21 de out. de 2026
After login: per-operation authorization, revocation and recovery for agent tools
An architecture exercise on tool authorization, revocation windows and interrupted remediation. Includes a local Node.js 22.13+ lab using built-in SQLite, without additional runtime dependencies or an AWS account.
20 de out. de 2026
Four engineering patterns that separate multi-agent from prompt chains
An AWS design exercise for agent workflows: bounded MCP tools, durable events, shared fallback validation and cost-aware routing, with a failure rehearsal and human approval before production changes.
07 de out. de 2026
Private CA logs every issuance: a field guide to IssueCertificateDetails
AWS Private CA’s new CloudTrail event carries the certificate details, requester and failure reason. A practical PKI audit playbook with EventBridge, Athena and retention controls.
03 de out. de 2026
Security Hub remediation plans: root-cause remediation, dissected
On October 1, 2026, Security Hub gained remediation plans: instead of a queue of findings, one plan per root cause, with an impact assessment and instructions in CLI, Terraform, CloudFormation, Python, and CDK. I dissect the root-cause remediation pattern (the problem it solves, its anatomy, when to use it and when not) and the detail that deserves the most scrutiny: AI agents consuming plans through the API.
03 de out. de 2026
Amazon Bedrock AgentCore Harness: From Idea to Production-Grade Agent
AgentCore Harness reached GA in June 2026 as a managed abstraction that collapses the LLM agent control plane into two API calls. In this article, I analyze how the harness works internally, where it fails, and what architects of financial-grade systems need to understand before putting it into production.