total-recall:llm-setup

total-recall LLM setup

Runs the operator-facing setup script that installs ollama and pulls the configured model (default qwen3.5:2b). After it succeeds the refinement layer is active and the SessionStart hook stops emitting the "not installed" notice.

Steps:

1. Run ${CLAUDE_PLUGIN_ROOT}/scripts/llm-setup.sh and stream its output to the operator. The script is idempotent — re-running on a fully-set-up machine is a no-op.
2. When it exits 0, confirm to the operator that the refinement layer is enabled and recommend running /total-recall:recall-rebuild so the next rebuild picks up the LLM refinement passes.
3. If it exits non-zero, surface the error verbatim and point at the relevant line in docs/llm-refinement.md.