Forward Deployed, Episode 2: Claude Code Skills and the Progressive Disclosure Problem

8 snips

Nov 17, 2025

Discover the game-changing Claude Code skills that are reshaping AI development. Noah reveals an efficient hook-based routing system that processes 3,000 tokens per second, bridging user messages with skills seamlessly. The debate over tiered action spaces highlights the differences between functions and skills. Explore how Andreessen's vision positions AI as a revolutionary computer architecture. Plus, innovative Git workflows eliminate merge conflicts, making collaboration smoother than ever.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Skills Are Progressive Context Folders

Claude Code skills are simple folders of prompts that let you extend model capabilities without bloating the system prompt.
Progressive disclosure lets the model discover capabilities on demand instead of loading everything upfront.

ANECDOTE

Alephic CLI Uses Fast Model As Router

Noah built an Alephic CLI with a skills repo and an MCP that searches skills on every message.
He layers a fast Cerebras Llama 120B model to write search queries and inject matched skills into prompts in about a second.

ADVICE

Keep Tier-One Tools Minimal

Do limit tier-one functions to a small set of atomic tools and push many actions out to a skills directory.
Use native file/search/bash tools so the model can find and run scripts without bloating system prompts.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Welcome back to episode 2 of Forward Deployed. This week Noah and Lance dive deep into Claude Code skills, a deceptively simple feature that’s changing how we think about building AI agents. Noah walks through the Alephic CLI he’s been building and reveals a wild hook-based routing system using Cerebras at 3,000 tokens per second.

Plus: Why Andreessen thinks AI isn’t the Internet redux.

Key Topics Covered

* Claude Code skills and the progressive disclosure problem

* How Noah built a hook to solve the 10% skill hit rate

* Tier 1 vs Tier 2 action space: Why Manus and Anthropic converged on the same architecture independently

* 3,000 tokens per second: Using Cerebras Llama 120B as an invisible routing layer for every user message

* The MCP/skill/command convergence: Are they all just different flavors of the same primitive?

* Vision feedback loop: Turning Gemini into a Pentagram creative director to critique Claude’s web designs

* Andreessen’s “computers v2” thesis: Why AI isn’t the Internet redux, it’s the first von Neumann architecture replacement in 80 years

* Git workflows with Claude Code: Why Lance and Noah don’t worry about merge conflicts anymore

Timestamps

* 00:11 – Welcome to episode 2 on Claude Code skills

* 00:28 – What are Claude Code skills? Not much more than a folder full of prompts

* 01:10 – Lance: Skills as “instructing a new hire” with subfolder instructions

* 01:25 – Simon Willison and Jesse Vincent’s “superpowers” discovery

* 04:45 – Noah demos the Alephic CLI skill directory structure

* 07:32 – The hook-based skill search system using Cerebras

* 08:19 – Lance reveals: YAML front matter always loads into system prompt

* 09:32 – The 10% skill hit rate problem when you have 10+ skills

* 10:08 – Cerebras Llama 120B running at 3,000 tokens per second for invisible routing

* 13:17 – The universal pattern: Everyone’s trying to control context

* 15:59 – Tier 1 vs Tier 2 action space: Manus and Anthropic converge independently

* 21:29 – Noah’s big challenge: Getting models to consistently look for skills

* 28:04 – Hit rate drops to 10% even with only 4–5 skills

* 30:54 – Could progressive disclosure become built-in like chain of thought?

* 34:06 – Lance on externalizing context to file systems

* 35:15 – Vision feedback loop: Gemini as Pentagram creative director critiquing Claude’s designs

* 37:57 – Andreessen: AI isn’t the Internet, it’s computers v2

* 42:08 – Why Noah and Lance don’t worry about merge conflicts anymore

Links & References

Core References

* 📄 Anthropic Engineering Blog on Skills – The engineering blog post Lance mentions about YAML front matter and system prompts

* 🎙️ Andrej Karpathy on Dwarkesh Podcast – Discussion on small models and reasoning engines

* 🎙️ Andreessen on Cheeky Pint Podcast – “AI is computers v2” thesis: First von Neumann architecture replacement in 80 years

* 🎙️ Claude Code Podcast with Boris and Kat – Discussion on dual-use tools

Tools & Frameworks

* 🔗 Claude Code – Anthropic’s AI coding assistant

* 🔗 Jesse Vincent’s Superpowers – Original skills plugin that inspired Noah

* 🔗 Manus – Consumer agent with multi-tier action space architecture

* 🔗 Cerebras – Llama 120B at 3,000 tokens per second

* 🔗 Puppeteer MCP – Browser automation MCP

* 🔗 Playwright MCP – Browser automation alternative

* 🔗 GitHub CLI – Command line tool Lance loves for PR management

Blog Posts

* Simon Willison’s Blog – First place Noah saw Claude Code skills coverage

* Cloudflare TypeScript Type Definitions Technique – Alternative to classic MCP definitions

Companies Mentioned

* Anthropic – Claude Code creator

* LangChain – Where Lance is a founding engineer

* Alephic – Noah’s AI consulting company

* Pentagram – Design agency Noah used as creative director persona

Development Tools

* Obsidian – Note-taking use case for Claude Code

* Git Work Trees – How the team manages multi-branch development

About the Hosts

Noah Brier is co-founder of Alephic, an AI consulting company helping brands build custom AI systems. He writes about AI strategy and implementation.

Lance Martin is a founding engineer at LangChain, where he works on developer tools for building AI applications.

Connect with the Hosts

* Noah Brier: LinkedIn | X/Twitter

* Lance Martin: LinkedIn | X/Twitter

Subscribe for weekly episodes exploring how AI is actually being deployed in the real world.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.forwarddeployed.com