From Total Recall
One-time setup for total-recall's optional local-LLM refinement layer — installs ollama (if missing) and pulls the configured model. Triggered when the SessionStart hook reports ollama / model missing.
How this skill is triggered — by the user, by Claude, or both
Slash command
/total-recall:llm-setupThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Runs the operator-facing setup script that installs ollama and pulls the configured
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:
${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./total-recall:recall-rebuild so the next rebuild
picks up the LLM refinement passes.docs/llm-refinement.md.npx claudepluginhub 88plug/claude-code-plugins --plugin total-recallRequests code review by dispatching a subagent with git diff context. Use after completing tasks, major features, or before merging to catch issues early.