Guilherme Salgueiro
Builder and AI systems practitioner. Guilherme helps developers, PMs, and founders move beyond prompting into structured AI system design — building with Claude Code, agents, and automation pipelines to ship products faster and more reliably.
62 guides by Guilherme Salgueiro
AI agent pipeline failure case study exercise
This exercise gives you a working three agent pipeline that just sent a false churn alert to an account manager, who...
Andre Albuquerque & Guilherme SalgueiroClaude Code operating model practice exercise
This exercise starts with a broken deploy that shipped because nobody defined what 'done' meant for an AI coding agent...
Andre Albuquerque & Guilherme SalgueiroClaude Code vs Cursor
Claude Code is a terminal-based agent built exclusively around Anthropic's own Claude models, while Cursor is a full...
Andre Albuquerque & Guilherme SalgueiroWhat Is Agent Memory in AI Systems?
Agent memory is an AI system's ability to store, retain, and recall information across sessions, turning a system that...
Andre Albuquerque & Guilherme SalgueiroWhat Is Agentic AI Product Management?
Agentic AI product management is the practice of scoping, designing, and shipping AI features built around agents...
Andre Albuquerque & Guilherme SalgueiroWhat Is an Agentic Workflow?
An agentic workflow is a process in which one or more AI agents interpret a goal, break it into steps, use tools, and...
Andre Albuquerque & Guilherme SalgueiroWhat Is a Skill in AI Coding Agents?
A Skill is a reusable, executable package of instructions, covering workflow steps, known pitfalls, quality gates, and...
Andre Albuquerque & Guilherme SalgueiroWhat Is an AI Agent?
An AI agent is software that uses a generative AI model to interpret a goal, decide on a course of action, and carry it...
Andre Albuquerque & Guilherme SalgueiroWhat Are AI Guardrails?
AI guardrails are the safeguards, spanning data, model behavior, and workflow design, that keep an AI system operating...
Andre Albuquerque & Guilherme SalgueiroWhat Are Hooks in Claude Code?
Hooks in Claude Code are shell commands that run automatically at specific points in a coding session, before or after...
Andre Albuquerque & Guilherme SalgueiroWhat Is Context Architecture for AI Agents?
Context architecture is the deliberate design of what information an AI agent has access to at every step...
Andre Albuquerque & Guilherme SalgueiroWhat Is a Context Window?
A context window is the maximum amount of text, measured in tokens, that a language model can process at one time...
Andre Albuquerque & Guilherme SalgueiroWhat Is Human in the Loop AI?
Human in the loop describes a system design where a person actively reviews, approves, or corrects an AI system's...
Andre Albuquerque & Guilherme SalgueiroWhat Is Multi-Agent Orchestration?
Multi-agent orchestration is the coordination of multiple specialized AI agents working together on a shared task, with...
Andre Albuquerque & Guilherme SalgueiroWhat Are Parallel Worktrees in Claude Code?
Parallel worktrees let a developer or AI coding agent check out multiple branches of the same Git repository into...
Andre Albuquerque & Guilherme SalgueiroWhat Is a Product Requirement Prompt?
A Product Requirement Prompt (PRP) is a structured document that combines a product requirement, curated context about...
Andre Albuquerque & Guilherme SalgueiroWhat Is a Quality Gate in an AI Pipeline?
A quality gate is an automated checkpoint in a pipeline that verifies a specific, measurable condition, tests passing...
Andre Albuquerque & Guilherme SalgueiroHow to Estimate LLM API Costs for a Product Feature
Estimate an LLM feature's monthly cost as requests per month x (input tokens x input price + output tokens x output...
Andre Albuquerque +2How to turn a PRD into agent tasks
Turning a PRD into agent tasks means cutting at seams the agent can verify, then ordering the cuts so nothing gets...
Andre Albuquerque & Guilherme SalgueiroGuilherme Salgueiro
Guilherme Salgueiro directs Building with Claude Code, Builders Camp's bootcamp on moving from prompting to production...
Andre Albuquerque & Guilherme SalgueiroAgent ready spec checklist for product managers
Ten checks decide whether a spec is executable or still ambiguous, and the two most commonly failed are the...
Andre Albuquerque & Guilherme SalgueiroHow to write a PRD for Claude Code
A PRD for Claude Code is written to be executed rather than discussed, so it carries four things a team spec does not...
Andre Albuquerque & Guilherme SalgueiroWhen to use an AI agent vs a workflow
Use a workflow when you can draw every step before the first input arrives, and an agent only when the path depends on...
Andre Albuquerque & Guilherme SalgueiroAI coding agent terms product managers need to know
Four terms out of this whole vocabulary change a product manager's decisions: context window, tool use, guardrails and...
Andre Albuquerque & Guilherme SalgueiroHow to use AI for product documentation that survives the next release
Documentation fails on maintenance, not on drafting, so point AI at drift detection before you point it at writing...
Andre Albuquerque +2How to use AI for product ops work that repeats every week
Product ops work splits cleanly into two piles, and only one is automatable: recurring artefacts with a right answer...
Andre Albuquerque +2Best AI agents for product managers in 2026
Claude Cowork is the best AI agent for most product managers in 2026: give it a goal and, per Anthropic's Cowork page...
Andre Albuquerque +2Best Claude Code skills for product managers in 2026
The best Claude Code skill for most product managers is Anthropic's free product-management plugin, which adds spec...
Andre Albuquerque & Guilherme SalgueiroBest MCP servers for product managers, chosen by the job you need done
Choose an MCP server by the product job it does and by whether it offers a read-only path, not by how often it appears...
Andre Albuquerque & Guilherme SalgueiroHow to build an AI agent without code
Build your first no-code AI agent from four n8n nodes: a Chat Trigger, an AI Agent, a chat model and Simple Memory...
Andre Albuquerque & Guilherme SalgueiroHow to Build an AI Assistant with MCP
Building an AI assistant with MCP means adding one or more MCP servers to an AI tool like Claude Code, scoping each...
Andre Albuquerque & Guilherme SalgueiroHow to run a competitor teardown with Claude Code without inventing facts
A competitor teardown in Claude Code is worth running only if every factual row can name the saved source file it came...
Andre Albuquerque +2How to manage context in Claude Code when you are a product manager
A Claude Code session degrades once the transcript fills with superseded decisions, and the fix is four habits: check...
Andre Albuquerque +2How to pull product data without SQL using Claude Code
Profile the dataset before you query it: column types, distinct values in every small column, null counts, and the date...
Andre Albuquerque +2Claude Code first week for product managers: a day by day ramp
Five days, 30 to 45 minutes each, ending with one artefact a colleague actually read. The ramp works because each day...
Andre Albuquerque +2Claude Code hooks: what they enforce that CLAUDE.md cannot, and how to use them in a product workflow
Claude Code hooks are commands Claude Code runs at a fixed lifecycle event, and a PreToolUse hook that exits with code...
Andre Albuquerque & Guilherme SalgueiroClaude Code permissions and safety for product teams
Permission rules are evaluated deny, then ask, then allow, and a deny at any level cannot be overridden anywhere else...
Andre Albuquerque +2Claude Code plugins for product managers, and how to ship one to your team
A Claude Code plugin bundles skills, agents, hooks and MCP servers into one installable directory with a manifest, so a...
Andre Albuquerque & Guilherme SalgueiroHow to verify AI-generated code as a product manager, before it merges or reaches a decision
Verify AI-generated code by asking for evidence instead of a summary: the test output, the build result, a screenshot...
Andre Albuquerque +2Claude Code scheduled tasks for product managers: cloud, Desktop or /loop
Use a cloud task for recurring work that must run without your machine, a Desktop task when the job needs your local...
Guilherme SalgueiroClaude Code setup for product managers: install to first useful output
A working Claude Code setup takes one install command, one login and one project folder, and Anthropic's quickstart...
Andre Albuquerque +2How to build Claude Code skills for product managers
Convert a prompt into a skill on its third use, not its first. A skill is a folder holding a SKILL.md file with a name...
Andre Albuquerque & Guilherme SalgueiroHow to build Claude Code slash commands for PM work
A slash command is one markdown file whose filename becomes the command, so .claude/commands/standup.md becomes...
Andre Albuquerque & Guilherme SalgueiroHow to use Claude Code subagents for PM workflows
A subagent runs one slice of a job in its own context window and hands back only the summary, which is why a four-hour...
Andre Albuquerque & Guilherme SalgueiroClaude Code use cases for product managers, and what each one is good for
Twelve Claude Code jobs are worth a product manager's time, and only five of them need repository access at all. The...
Andre Albuquerque +2Claude Code vs GitHub Copilot for product managers: agent work or completion
GitHub Copilot began as completion inside a developer's editor and added an agent mode; Claude Code began as a terminal...
Andre Albuquerque +2How to use parallel worktrees in Claude Code
Running claude --worktree twice with different names gives you two isolated checkouts of the same repository, each on...
Andre Albuquerque +2How to write a CLAUDE.md file as a product manager
Keep CLAUDE.md under 200 lines and fill it only with rules that are true in every session on this project, which for a...
Andre Albuquerque & Guilherme SalgueiroClaude Code CLAUDE.md vs skills vs hooks vs subagents: where each rule belongs
Put a rule in CLAUDE.md if Claude must always know it, in a skill if it is needed only for some tasks, in a hook if it...
Andre Albuquerque & Guilherme SalgueiroHow to use Codex to build a feature step by step: plan, build in slices, ship
Use Codex to build a feature in five steps: ask for a plan in Plan mode, edit the plan until every slice names its...
Andre Albuquerque & Guilherme SalgueiroHow to use Codex app for non developers: setup, plans and a safe first task
To use the Codex app as a non-developer, install the ChatGPT desktop app (macOS, Windows, or Linux in preview), sign...
Andre Albuquerque & Guilherme SalgueiroHow to create a skill in Codex: turning a repeated instruction into a reusable workflow
Create a Codex skill on the third time you type the same instructions: a folder in .agents/skills holding a SKILL.md...
Andre Albuquerque & Guilherme SalgueiroOpenAI Codex vs Claude Code for non-engineers: where each runs and how each asks permission
Both tools run in a terminal, a desktop app and the cloud, and both ask permission before risky actions, so the choice...
Guilherme SalgueiroHow product managers work with engineers using AI agents
The line between product and engineering work stopped being technical skill and became ownership of consequence: build...
Andre Albuquerque +2How to Design an AI Agent
Designing an AI agent means deciding what it can see, what it may use, when a human has to approve its action, and what...
Andre Albuquerque & Guilherme SalgueiroHow to write AGENTS.md for Codex: a 25-line file for a product repository
Write AGENTS.md for Codex as a short contract: what the repo is, the exact build, test and lint commands, the few rules...
Andre Albuquerque & Guilherme SalgueiroMCP for product managers, and what the protocol actually changes
MCP is an open standard, announced by Anthropic on 25 November 2024, for connecting an AI application to external...
Andre Albuquerque & Guilherme SalgueiroMCP setup for product managers, from first connection to first useful answer
You can connect an MCP server in the Claude apps with a URL and no command line, on any plan from Free to Enterprise...
Andre Albuquerque & Guilherme SalgueiroMCP troubleshooting for product managers, five failures and how to tell them apart
Almost every broken MCP connection is one of five things: it never connected, it is not authenticated, it is registered...
Andre Albuquerque & Guilherme SalgueiroMCP vs plugins and integrations, and when a protocol beats a point connection
A protocol wins when many clients need many tools, because one server serves every client that speaks MCP instead of...
Andre Albuquerque & Guilherme SalgueiroHow to review Codex pull requests: evidence, scope and the merge decision
Review a Codex pull request in three passes: check the evidence the description carries, compare the description's file...
Andre Albuquerque & Guilherme SalgueiroWhat is agentic coding for product managers, and what changes for you
Agentic coding is the shift from a model that completes your next line to one that explores a codebase, plans an...
Andre Albuquerque & Guilherme Salgueiro