At a glance
WHAT IT’S REALLY ABOUT
Advanced Claude Code for PMs: MCPs, skills, workflows automation
- Claude Code’s advantage is depth for “power users,” enabling long-running, tool-connected workflows that replace constant context switching across apps.
- MCP (Model Context Protocol) is presented as the standard way to connect Claude Code to work systems (Docs, Linear, Slack, Notion, etc.) to pull and push data without copy/paste.
- Skills are reusable, semi-automatic workflow recipes (from repos, marketplaces/plugins, or custom-built) that standardize outputs like PRDs, editing, research synthesis, and document creation.
- The episode demos an end-to-end PM workflow: draft a survey, analyze responses, write a PRD, generate a deck (Google Slides), and create prioritized engineering tickets in Linear.
- GitHub integration turns Claude into a “remote worker,” letting you @Claude in issues to make repo changes even when you’re away, with notes on basic git concepts (pull/push/branches).
IDEAS WORTH REMEMBERING
5 ideasClaude Code wins by enabling deep, repeatable workflows—not one-off prompts.
The key productivity jump comes when you notice recurring work, encode it as files/rules/skills, and then run the workflow “in one shot” instead of step-by-step prompting.
Start by connecting your document system via MCP for immediate PM ROI.
Because PM work is largely document-driven, hooking up Google Drive/Docs, Notion, or Confluence first makes it easy to draft in Claude Code and publish where stakeholders already collaborate.
MCPs provide “tooling for free,” while APIs require you to build and maintain glue code.
MCP servers expose many pre-defined actions (e.g., Linear’s many issue/project tools) that Claude can use immediately; APIs (e.g., Gemini image generation) typically require scripts, keys, and documentation context.
Turn on only the MCPs you need to preserve context window capacity.
Each enabled MCP consumes some context, so the practical pattern is to keep a small active set (e.g., Linear + Google) and enable others (e.g., Reddit) only when needed.
Skills are the lever for consistent quality—but don’t rely on auto-triggering yet.
Skills can encode templates and “house style” for PRDs, editing, or research synthesis, but the demo shows triggers can be unreliable; explicitly instructing Claude to use a specific skill is a safer current practice.
WORDS WORTH SAVING
5 quotesI spend all day in Claude Code because it can do more and more parts of my workflow. It gets to the point where, as a product manager or anyone working with these tools, you never really have to even leave Claude Code.
— Carl Vellotti
Opus 4.5 is basically AGI, where you can give it a task, and you can kind of just trust it to, like, walk away from your computer, and it will just keep going forward in the task.
— Carl Vellotti
I think that's really what it is, is it's just thinking about the whole system that you're building with Claude rather than just having it do individual tasks.
— Carl Vellotti
Creating tickets as a product manager depends on how your team is set up. Sometimes engineers do it, sometimes PMs do it. But just in general, it's always, like, such a painful process, 'cause you know the information, and you just have to, you know, go through the UI and click all these buttons and, like, kind of make sure everything's all organized. Whereas now with Claude, it can just literally create all those tickets for you.
— Carl Vellotti
Any time you're starting, like, a new task or, or something else, just see, "Can I do it in Claude Code?" is my best advice. Just, "Can I do it in Claude Code?"
— Carl Vellotti
High quality AI-generated summary created from speaker-labeled transcript.
