README written by a human for humans
This is a VSCode extension for viewing the RESOLVED state for your agent, like Claude Code. Once a harness is initialized, there's a merger of skills, system prompt mds, memories etc. This extension provides a visualization of all these, in a way that makes maintaining agents easier.
Current State
- Skills View, coming from the user specific
~/.claude/folder, the project's global.claude/folder, and active plugins. - System Prompt View, merging CLAUDE.md or similar files.
- Memory View, for purging / tracking stale or important memories
- Hooks for tool invocations etc
- Permissions
URI Support
(Basically the reason why I'm doing this project) The extension comes with URI support for opening the views, specific skills, and hopefully in the future, actual useful actions. The core idea here is using Extension URI support as a better, progressively disclosed (and thus, scalable) MCP Server. A minimal intent indication can point towards a --help .md file which indicates the best ways to have your Agent interact (open, update) this UI. This makes VSCode an App Marketplace with built-in agent interactions (and BYOA!!).
vscode://canoq.claude-viewer/skill/commit
vscode://canoq.claude-viewer/skill/deploy?scope=user # a specific scope's copy
vscode://canoq.claude-viewer/skill # just open the picker
Install
Ok, from this point on, this part was AI generated, it's boring
Search for Claude Viewer in the Extensions view, or:
code --install-extension canoq.claude-viewer
Then open the Command Palette and run Claude Viewer: Open.
Unofficial
This is a community project. It is not affiliated with, endorsed by, or supported by Anthropic. "Claude" and "Claude Code" are trademarks of Anthropic, used here only to describe what the extension reads.
License
MIT