Problem

When running ck inside a subdirectory of an already indexed project, a new .ck directory is created in that subdirectory. This leads to redundant indexes, increased disk usage, and unnecessary re-indexing work.

Proposed Behavior

When executed in a subdirectory, ck should:

1. Traverse upward from the current working directory to locate an existing .ck directory.
2. If found, reuse that index instead of creating a new one.
3. Scope search results to files within the current subdirectory only.

Example

Given:

project/
.ck/
src/
module/

Running:

cd project/src/module
ck “query”

Expected behavior:

• Use project/.ck as the index
• Return only results under project/src/module

Optional Optimization

For performance, ck could optionally:

• Prioritize scanning and updating only files within the current subdirectory
• Defer or avoid reprocessing unrelated parts of the repository

Benefits

• Avoids duplicate .ck directories
• Reduces disk usage and indexing time
• Improves usability in nested workflows
• Aligns with common tooling patterns (e.g., git, npm, etc.)

Notes / Open Questions

• Should there be a flag to disable parent lookup (e.g., --no-parent-index)?

Thanks so much for building such a wonderful tool!