rag-exportlisted
Install: claude install-skill noory-code/noory-ai
# rag-export — making a snapshot
> Before executing this workflow, read and apply `../HOST_CONTRACT.md`.
## Steps
1. **Prerequisite check**: `rag_get_settings`.
2. **Determine output path**: extract path from the user's utterance. If none, default to `~/Downloads/rag-snapshot-<YYYYMMDD-HHMM>.tar.gz`.
- If only a directory is given, auto-name the file inside it with the pattern above.
- If a file with the same name already exists, confirm whether to overwrite.
3. **Call**: `rag_export(out_path=<resolved>)`.
- The response includes header metadata (embedding model/dim/stats).
4. **Output the result**:
```
📦 Snapshot created
file: <out_path>
size: <du -h>
embedding: <model> (dim <N>)
stats: files M, chunks N, entities P, ...
```
5. **Share guidance**: "Send this file to a teammate and they can receive it with `/rag:rag-import <path>`. For more detailed share scenarios: `/rag:rag-share-guide`."
## Error handling
- Not enough disk space, no write permission → report clearly.
- Even if the index is empty, export is possible (an empty snapshot with header only).