xllm-setup

/xllm-setup — Install health check

<state> = project state dir: .xllm/ (default) or legacy .grok/ when the project already uses it.

Run these steps (do not only describe them).

1. Resolve + remember advisor path

Find xllm-advisor.js via (first hit):

• .xllm/xllm-advisor-path (legacy: .grok/xllm-advisor-path)
• XLLM_ADVISOR_PATH / GROK_PLUGIN_ROOT / XLLM_PLUGIN_ROOT
• ./scripts/xllm-advisor.js
• plugin checkout path the user has open

Then run:

node <advisor.js> --remember
node <advisor.js> --which

Confirm .xllm/xllm-advisor-path (or legacy .grok/) exists and points at a real file.

2. Artifact directories

Ensure:

<state>/artifacts/ask|xllm|proposals|exec/

(--remember / --doctor also create these.)

2.5 Project advisor wizard (posture packs — optional)

Resolve pins deterministically; the skill only renders and confirms.

1. Preview the recommended pack (default balanced):

   node <advisor.js> --setup balanced --json
   

   The resolver returns { roles, warnings, evidence, recommended_packs }. balanced leaves analysis/design OPEN (measured routing) and pins at most a free local critic — a pin FREEZES measured routing, so packs pin only genuine constraints. quality = max-spend lock, frugal = cost lock, local = offline lock, skip = clear pins.

2. Ask ONE question, offering the first four of recommended_packs (always include skip). Show the effort legend (Quick=low / Standard=medium / Deep=high) and one-line role glosses. Never invent cloud model names — cloud pins omit the model.

3. Show the resolved preview (roles + warnings + which stay OPEN and why), then on the user's accept:

   node <advisor.js> --setup <pack> --apply
   

   Partial tweak: --role analysis=grok@high (validated; one bad override writes nothing). Reverting: node <advisor.js> --setup skip --apply clears the posture pins. Verify with node <advisor.js> --profile-show.