Close a work arc — write the next docs/history/NNNN-<arc>.md continuity entry, update the history index and mkdocs nav, strike completed items in docs/wip/NEXT_STEPS.md, append any D-NNNN operational decisions to 0002-decisions.md, and suggest the version tag. Use when wrapping up a multi-step session or a milestone. Keeps the strict docs build green.
Pre-deploy sanity check for Docker Compose stacks. Validates each compose file with `docker compose config` (read-only), then flags undefined interpolation variables and missing .env before the stack is brought up. Use before `docker compose up`, when editing a compose file, or whenever asked to validate or check a compose stack. Never brings stacks up or down.
One-shot read-only health snapshot of the DGX host — GPU mode, GPU utilization and compute-app count, nvidia-smi memory and processes, and running containers. Use to answer "what is the DGX doing right now" before deploying, switching GPU mode, or debugging a stuck endpoint. Strictly read-only; never changes state.
Validate that the docs site builds cleanly in strict mode before committing or pushing docs changes. Detects the project's docs toolchain (Makefile docs target, MkDocs, Sphinx, Docusaurus, mdBook) and runs its strict build, then reports PASS/FAIL with the exact unresolved refs or warnings. Use when about to commit or push changes under docs/, mkdocs.yml, or docs config, or whenever asked to check or validate the docs build.
Show per-subagent invocation counts and token/cost spread by model tier from the current Claude Code transcript — proves whether the model-tiered fleet (haiku/sonnet/opus) actually spreads token cost. Use to see who was invoked, how often, tokens per agent/model, and escalation rate. Local and read-only (no Grafana/OTEL needed).
Coordinate which vLLM owns the single DGX GPU before using a local vLLM endpoint. Checks the active mode (code/research/idle/BROKEN-BOTH) read-only, and switches only on explicit request. Use before pointing any tool at a local vLLM (http://<dgx>:9000/v1 or :8003/v1), when a vLLM call fails to connect, or when asked to bring the coder-next or autoresearch vLLM up or down. Switching is shared-state — never switch without explicit approval.
Resolve a homelab service name to its real endpoint URL and how to call it — DGX host, vLLM coder-next/autoresearch slots, Ollama, observability labels. Use when an agent or client needs to point at a homelab service, when writing config that consumes one, or when asked "what's the URL for X". Read-only.
Scaffold a new documentation file — ADR, RFC, recipe, or wip note — from the repo's conventions, assign the next number/slug, wire it into the mkdocs nav, update the section index, and keep the strict build green. Use when starting an ADR/RFC/recipe/design note, or when asked to write up a decision or a runbook.