Claude just got another superpower (but they kept it quiet!)
A creator discovers buried features in a recent Claude desktop app update, including file system visibility, split-view sessions, and markdown rendering. These changes make the desktop app a viable replacement for terminal-based Claude Code workflows. The creator argues this enables structured, context-aware agentic project management without needing VS Code or a terminal.
Summary
The video begins with the creator expressing surprise at needing to make the video, prompted by a Claude desktop app update that was announced with a focus on parallel sessions and in-app review — but which buried two features the creator considers far more significant.
The creator provides context on the evolution of Claude's desktop experience, noting that prior to this update, the app was essentially a one-dimensional chat interface. There was no visibility into a broader business context, no connection between sessions, and no ability to track progress across projects. This limitation was the reason the creator's team built their own 'agentic operating system' — a pre-built file structure that stores context, loads it selectively, and improves outputs over time. The problem was that this system required terminal access because the desktop app couldn't see or interact with the underlying files.
The first major buried feature highlighted is the new file system sidebar. Claude's desktop app can now display and interact with the underlying file structure, including project folders, output files, and configuration files like CLAUDE.md. This means users can navigate into client-specific folders, view and edit context files (brand voice, positioning, agent rules), and work within a structured agentic OS directly from the desktop UI — something previously only possible via terminal or VS Code.
The second major feature is the ability to view and render output files in markdown within the app itself. Users can now open output files stored in project folders and see them rendered in a formatted markdown preview, with the ability to directly edit them. The creator notes this is significant because it removes the need to switch to external tools just to read or modify outputs.
Additional features covered include a multi-session split view (up to four simultaneous independent sessions), a new plan sidebar that renders Claude's session plans in formatted markdown alongside the chat, and the ability to select which folder to start a new Claude Code session in — enabling scoped, context-rich work per client or project.
The creator also outlines three current limitations: PNG and non-markdown files aren't rendered visually (they display as code instead), hidden skills folders are not yet accessible or editable within the desktop app, and .env files containing API credentials cannot be viewed or modified through the desktop UI (though Claude can still access and use them for API calls). The creator expects these to be addressed in future updates.
The video closes with the creator stating they can now see themselves fully moving away from VS Code for Claude-based workflows, and inviting viewers to share what features might still be holding them back.
Key Insights
- The creator argues that the most impactful features of the Claude desktop update were not highlighted in Anthropic's official announcement, which focused on parallel sessions and in-app review rather than file system access and markdown rendering.
- The creator claims their team built a custom 'agentic operating system' — a structured file system for storing and selectively loading context — specifically because the Claude desktop app previously had no visibility into underlying files, forcing terminal use.
- The new file sidebar allows users to navigate their project folder structure, view and edit files like CLAUDE.md, access client-specific brand context, and render output files in markdown — all without leaving the desktop app.
- The creator notes that users can now choose which folder to start a Claude Code session in, enabling scoped, context-aware work within individual client or project directories rather than operating in an isolated, context-free environment.
- The creator identifies three current limitations: non-markdown files like PNGs render as code rather than visual previews, hidden skills folders are not accessible or editable in the desktop UI, and .env credential files cannot be viewed or changed through the app despite Claude still being able to use them.
Topics
Transcript
[0:00] Holy smokes. I did not expect to be making this video today. But if you've been using claw code in the terminal inside VS Code or anything like that and you've been waiting for a reason to switch to the desktop app, then today might be that day. So last week Claude dropped an update and while the announcement focused on running sessions in parallel and reviewing things without leaving the app, they forgot to mention or they buried two features that are actually going to change how most people use Claude inside the desktop. So the real issue before was that the desktop [0:30] app was always super one-dimensional. So remember back to the days of chatbt chatting…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Simon Scrapes
I Rebuilt Hermes in Claude Code (It’s Ridiculously Good)
The creator rebuilds key features of the Hermes agentic AI system inside their own Claude Code setup rather than installing it off-the-shelf, arguing that understanding the architecture provides more long-term leverage. They identify three hidden costs of pre-built systems: inherited assumptions, inability to fix unknown code, and poor scalability. Their custom build addresses Hermes's limitations in multi-client identity management, semantic memory recall, and skill system maintainability.
Agentic AI Systems, Clearly Explained
This video explains agentic AI systems across four levels of complexity: chatbots, AI workflows, agentic workflows, and agentic AI systems. Using content repurposing as a consistent example, the speaker demystifies intimidating terms like harnesses, MCPs, and memory systems, arguing that these systems are fundamentally just files and folders accessible to non-developers.
Stop Downloading Claude Code Skills. Do This Instead.
The video argues that downloaded Claude Code skills are too isolated or bloated to drive real business outcomes. The speaker introduces 'skill systems' — orchestrated chains of small, focused skills that feed into each other end-to-end. A practical example shows a five-skill automation that converts a YouTube video into five ready-to-post short-form clips.
Master 80% of Claude Code, Just Learn These 15 Things
This tutorial covers 15 foundational concepts for using Claude Code effectively, emphasizing that context management is the single most critical skill. The presenter draws from real user behavior across thousands of users to show how plan mode, context tools, slash commands, skills, hooks, MCPs, and memory systems combine to form a powerful agentic workflow.
Master Claude Design in One Video (Complete Build)
The video demonstrates a live build using Claude Design to create real business assets for the Aentic Academy, including a landing page, mobile wireframe, and slide deck. The host emphasizes that building a proper design system upfront is the single most important factor in avoiding generic AI aesthetics. The video compares multiple methods: using ChatGPT's image model, the Skill UI CLI tool, and Claude Design's native interface.