[FEATURE] Claude Desktop: Support for working is subfolder of mono-repo

Resolved 💬 3 comments Opened Apr 7, 2026 by thedug Closed May 25, 2026

Preflight Checklist

  • [x] I have searched existing requests and this feature hasn't been requested yet
  • [x] This is a single feature request (not multiple features)

Problem Statement

Title: Mono-repo support: session list doesn't reflect sub-project context

Problem:

Claude Desktop doesn't surface which subdirectory a session belongs to, making mono-repo workflows unmanageable. All sessions appear in one flat list with auto-generated names, even when they were started from different sub-projects with different CLAUDE.md files and completely different contexts.

The data is already there. Every session record includes a cwd field. Desktop just doesn't use it.

Real-world example:

We run 6 AI agent personas out of one repo, each in its own subdirectory with its own CLAUDE.md, workspace, and session history:

vc/claude/
morgan/CLAUDE.md (operations)
nate/CLAUDE.md (sales)
lena/CLAUDE.md (marketing)
elise/CLAUDE.md (research)
sloane/CLAUDE.md (engineering)
kira/CLAUDE.md (personal)

On a typical day we have 15-20 active sessions across these agents. The sidebar is an undifferentiated wall of auto-generated names. Finding a specific session requires opening each one.

Proposed Solution

Show the sub-folder in session titles or somewhere pomiently in the session and session list If cwd is a subdirectory of the project root, prefix or tag the session name with the folder name (e.g. "morgan/: Weekly Brief Pipeline").
Filter/group the session list by sub-folder. Let users narrow the sidebar to sessions from a specific subdirectory.
Sub-project picker when starting a session. A configurable list of sub-projects (via launch.json, project config, or auto-detected from subdirectories with CLAUDE.md files) so users can quick-launch into a specific context instead of always landing at the repo root.
Any of these would make mono-repo workflows viable. Today the only workaround is manual renaming, which doesn't scale.

Alternative Solutions

_No response_

Priority

High - Significant impact on productivity

Feature Category

Other

Use Case Example

_No response_

Additional Context

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗