bycq27-dev· 4 skills
Use to enable and operate rag-rat's optional AI memory-maintenance passes — the dream --verify / --compact model turns that verify repo memories (flag claims that no longer hold) and compact them (shorter summaries). Off by default. This configures [llm.dream] / [llm.dream.remote] (a chat model, over a Connect endpoint or an ephemeral Modal/RunPod cookbook box — vllm/ollama, never infinity), and sets up running it on demand or on a schedule (systemd timer). Pairs with the dream-review skill, which triages the findings these passes produce. Triggers: "configure dream", "enable dream", "AI memory verification", "memory compaction", "schedule dream maintenance", "set up the dream model".
$npx -y skills add cq27-dev/rag-rat --skill configure-rag-rat-dreamInstalls into the current project.
Run `npx skills use "https://github.com/cq27-dev/rag-rat" --skill "cq27-dev/rag-rat/configure-rag-rat-dream"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/cq27-dev/rag-rat" that are relevant to the current task. Run `npx skills add "https://github.com/cq27-dev/rag-rat"` and select the relevant skills, then follow their instructions.