richfrem/vector-db-ingest

Dependencies

This skill requires the chromadb and langchain packages defined in the plugin root.

---

VDB Ingest Agent

Role

You ingest (index) repository files into the ChromaDB vector store so they can be semantically searched. You build or update the parent-child chunk structure that query.py searches against.

High-Performance Mode: This skill uses a configurable batch processing engine (default 1,000 files) defined in .agent/learning/vector_profiles.json.

Prerequisites

1. First-time setup

If vector_profiles.json is missing, run the init skill first:

python ./scripts/init.py

2. Execution Mode

This plugin defaults to In-Process mode for zero-latency direct disk access. No background server is required unless explicitly configured in the profile.

Execution Protocol

Full ingest (first time or full rebuild)

Note: The --profile flag is mandatory to load the correct manifest and batch settings.

python ./scripts/ingest.py --profile wiki --full

Incremental ingest (only new/changed files since N hours)

python ./scripts/ingest.py --profile wiki --since 24

Single File/Folder Ingest

python ./scripts/ingest.py --profile wiki --file path/to/file.md
python ./scripts/ingest.py --profile wiki --folder path/to/folder

After Ingesting

Run a quick semantic search to confirm the new content is retrievable:

python ./scripts/query.py "search query" --profile wiki --limit 3

Rules

• Profile Sovereignty: Always pass --profile to ensure the correct batch size and manifest are used.
• In-Process Reliability: Ensure no other process is holding a lock on the database folder during ingestion.
• Source Transparency: State which profile was ingested, how many files, and any errors encountered.