Cal Corum 51fe634ff5 refactor: convert 5 more skills to commands, update transcriber defaults

Convert backlog, project-plan, save-doc, youtube-transcriber, and
z-image from skills/ to commands/ so they appear as user-invocable
slash commands with plugin name prefixes.

Update youtube-transcriber: switch default model from gpt-4o-transcribe
to gpt-4o-mini-transcribe (OpenAI's current recommendation, half cost)
and fix cost estimates that were 4-7x too high.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-03-19 14:41:37 -05:00

1.4 KiB

Raw Permalink Blame History

title

description

type

domain

KB-RAG Reindex Runbook

When to Use

Search returns stale or missing results after a push
After bulk file additions or directory restructuring
After recovering from container crash or volume issue

Incremental Reindex (normal)

Only re-embeds files whose content hash changed. Fast (~1-5 seconds).

ssh manticore "docker exec md-kb-rag-kb-rag-1 md-kb-rag index"

Full Reindex (nuclear option)

Clears the state DB and re-embeds everything. Slow (~2-3 minutes for 150+ files).

ssh manticore "docker exec md-kb-rag-kb-rag-1 md-kb-rag index --full"

Verify

# Check health
ssh manticore "docker exec md-kb-rag-kb-rag-1 md-kb-rag health"

# Check indexed file count and Qdrant point count
ssh manticore "docker exec md-kb-rag-kb-rag-1 md-kb-rag status"

# Check for validation errors in recent logs
ssh manticore "docker logs md-kb-rag-kb-rag-1 --tail 30 2>&1 | grep WARN"

Escalation

If Qdrant won't start: check disk space on manticore (df -h)
If embeddings OOM: check GPU memory (ssh manticore "nvidia-smi")
Full stack restart: ssh manticore "cd ~/docker/md-kb-rag && docker compose down && docker compose up -d"

1.4 KiB Raw Permalink Blame History