Criador do bridgecode

Marcelo Santiago

Developer and creator of bridgecode, an agentic development environment for orchestrating squads of AI agents. Helps people going from vibe coding to professional development move from idea to shipped product.

GitHub Twitter / X LinkedIn Instagram Threads Website

Artigos publicados

252 artigos por Marcelo Santiago

Official image of Code Storage with a silver metallic figure, a black band with the text CODE.STORAGE and the signature PIERRE COMPUTER COMPANY.

Newsletter

Nov 4, 20266 min

OpenAI to acquire Pierre and expand Codex tools

OpenAI will acquire Pierre to expand Codex with Git and review tools. See what is already confirmed and what depends on closing.

#OpenAI#Pierre Computer Company#Codex

Marcelo Santiago Read

Dicas

Nov 4, 20266 min

5 tips for working with agents in VS Code 1.145

Use VS Code 1.145 to organize agent sessions, test remote apps, schedule local automations and review changes before integrating.

#VS Code#VS Code 1.145#agentes de IA

Marcelo Santiago Read

Four panel chart compares predicted and observed values in a QSAR model with scaffold split

Dicas

Nov 4, 20266 min

5 tips to organize agents with Google Antigravity

Learn from Google DeepMind how to separate research, deterministic calculations and human validation in an agent workflow in Google Antigravity.

#Google Antigravity#quimioinformática#fluxo de agentes

Marcelo Santiago Read

GitHub social card for the modelcontextprotocol/servers repository, with the MCP symbol, project name and repository indicators.

Dicas

Nov 4, 20266 min

mcp-server-git: MCP Git for AI CLIs

Meet the official MCP Git server: uvx installation, conditional use in AI CLIs, license and security risks verified on GitHub.

#MCP#mcp-server-git#Git

Marcelo Santiago Read

Official release card for Claude Code v2.1.318 on GitHub, with the version, the New Release title and the Anthropic icon.

Newsletter

Nov 4, 20265 min

Claude Code 2.1.318 reinforces credentials and MCP

Claude Code 2.1.318 surfaces login expiration, adjusts model forks and fixes issues in MCP, settings and sessions.

#Anthropic#Claude Code#v2.1.318

Marcelo Santiago Read

Newsletter

Nov 3, 20266 min

Claude Code 2.1.317 expands controls and notifications

November 2 release adds permission notices, fail-block hooks, cost reports and fixes for sessions and MCP.

#Anthropic#Claude Code#v2.1.317

Marcelo Santiago Read

Two adults read an illustrated book to a baby, in a scene tied to the everyday tasks evaluated by LifeVal.

Benchmarks

Nov 3, 20266 min

What does the LifeVal benchmark actually evaluate in real life?

OpenAI introduced LifeVal, an AI evaluation for everyday tasks like childcare, finances and public services. See its scope and limits.

#LifeVal#OpenAI#benchmark

Marcelo Santiago Read

Official thumbnail of the introductory Serena video, with the title in English and a play button.

Dicas

Nov 3, 20265 min

Serena MCP: how to give your agent semantic navigation in VS Code

Serena connects agents to code via MCP and offers symbol driven navigation, search and editing. See how to install it in VS Code, activate a project, and understand the security and license limits before you use it.

#Serena#MCP#VS Code

Marcelo Santiago Read

Thumbnail of the GitHub Spec Kit video, with the project name and an illustrated robotic character.

Dicas

Nov 3, 20265 min

GitHub Spec Kit: how to turn requirements into steps for AI agents

GitHub Spec Kit organizes agent-driven development into specifications, plans, tasks, implementation and verification. See the prerequisites, a short workflow and the cautions for adopting the tool in an existing repository.

#GitHub Spec Kit#Spec-Driven Development#agentes de programação

Marcelo Santiago Read

Official illustration of Chrome DevTools for agents: browser, developer tools and AI agent.

Dicas

Nov 3, 20265 min

Chrome DevTools MCP: how to debug web pages with an AI agent

Chrome DevTools MCP connects programming agents to Chrome's inspection tools to investigate console, network, accessibility and performance. See how to install it in VS Code, test it on a safe page and protect your browser data.

#Chrome DevTools MCP#MCP#agentes de programação

Marcelo Santiago Read

Figure from the FeatureLoop paper comparing three ways to develop with agents: full specification, progressive requests, and progressive requests with corrections.

Benchmarks

Nov 2, 20267 min

FeatureLoop: benchmark reveals failures of agents with evolving requirements

The FeatureLoop preprint evaluates code agents on tasks with evolving requirements. The study shows that feedback helps, but local fixes do not always preserve everything that was already requested.

#FeatureLoop#benchmark de agentes de código#requisitos em evolução

Marcelo Santiago Read

Screenshot of the official GitHub Agentic Workflows quick start guide, with prerequisites and steps to set up a daily automation in the repository.

Dicas

Nov 2, 20267 min

How to automate repository tasks with GitHub Agentic Workflows

Learn how to install GitHub Agentic Workflows, add a sample automation, and review permissions, authentication, and generated files before running it.

#GitHub Agentic Workflows#gh-aw#GitHub Actions

Marcelo Santiago Read

The biggest AI newsletter in the World

Follow the evolution of AI squads and multi-agent systems.

In-depth technical curation: agentic architectures, real token benchmarks, MCP servers and open-source tools in your inbox and on the blog.