Create Codebase CSV

First: mantis use get_space_context (unless you just ran it this turn).

Use mantis create codebase. This command is local-first: it scans files on the user's machine and writes a CSV. It can optionally continue into mantis create map.

CSV-only Flow

1. Get the repo root from $ARGUMENTS or ask the user.
2. Run:

mantis create codebase "./repo" --out "./repo-codebase.csv"

3. Explain the CSV columns:
   • path: repo-relative file path
   • file_name
   • extension
   • language
   • kind: rough file role such as component, route, api, utility, test, source
   • loc
   • bytes
   • imports
   • summary
   • content

Full Demo Flow

If the user wants the codebase to become a Mantis map immediately, run:

mantis create codebase "./repo" \
  --create-map \
  --space-mode new \
  --space-name "Codebase: Repo Name" \
  --private \
  --map-name "Source Code Index" \
  --activate \
  --thread-name "Architecture Review"

The command uses good defaults for codebase maps:

• title: path
• semantic: summary,content,imports
• categoric: language,kind,extension
• numeric: loc,bytes

After activation

If --activate was used, the new thread is active immediately. Explore with:

mantis use get_space_context

Example follow-up questions:

What are the main architectural zones in this codebase?
Which files are central or risky?
Where should I start if I want to refactor the map rendering system?

Notes

• Use mantis setup first if API configuration is missing.
• Avoid scanning dependency/build folders manually; the CLI already ignores common junk like .git, node_modules, .next, dist, build, and caches.
• For agent-driven demos, pass flags explicitly instead of relying on interactive prompts.