docs(prep-my-visit): add a wellness-guide README and clean ASCII text

Package a detailed skill README and remove non-ASCII glyphs from runtime instructions so docs stay portable, readable, and ready for every terminal.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: pswider

skills/prep-my-visit/README.md

@@ -0,0 +1,141 @@
+# prep-my-visit
+
+`prep-my-visit` prepares a patient for an upcoming visit with a concise, IPS-aligned package:
+
+- provider-facing summary
+- patient companion summary
+- pre-visit lab opportunities
+- portal-ready snippets (patient-reviewed before send)
+
+The goal is simple: start the visit with context and current data, not chart archaeology.
+
+## Why this skill exists
+
+Most pre-visit friction is predictable:
+
+- context is fragmented across portals and timelines
+- key labs are often ordered but not completed before the encounter
+- patients arrive without a clear goals list
+
+This skill turns those failure modes into a structured prep flow with explicit safety boundaries.
+
+## What it produces
+
+Artifacts are persisted under:
+
+`~/.openclaw/workspace/tula/briefs/{visit_id}/`
+
+Expected outputs:
+
+- `provider.pdf`
+- `patient.pdf`
+- `ips-bundle.json`
+- `lab-opportunities.json`
+- `snippets/*.txt`
+
+## Safety model
+
+`prep-my-visit` is constrained by design:
+
+- no diagnosis, prognosis, or treatment recommendations
+- no auto-ordering tests
+- no auto-send to patient portals
+- no insurance, billing, prior-auth, or EOB workflows
+- no PHI transfer outside `~/.openclaw/workspace/`
+
+## Skill structure
+
+- Runtime instructions: [`SKILL.md`](SKILL.md)
+- Operational workflow: [`references/workflow.md`](references/workflow.md)
+- Lab analyzer rules: [`references/lab-analyzer.md`](references/lab-analyzer.md)
+- IPS section contract: [`references/ips-contract.md`](references/ips-contract.md)
+- Script usage: [`references/scripts.md`](references/scripts.md)
+- Prompt examples: [`references/examples.md`](references/examples.md)
+
+## Local quality checks
+
+From repo root:
+
+```powershell
+waza check skills/prep-my-visit
+```
+
+Deterministic script checks (run inside skill context):
+
+```powershell
+node {baseDir}/scripts/validate_ips_sections.mjs <composition.json>
+node {baseDir}/scripts/validate_lab_opportunities.mjs <lab-opportunities.json>
+node {baseDir}/scripts/enforce_snippet_limits.mjs <snippets.json>
+```
+
+## Use with Claude
+
+Use this skill in Claude-backed runtime via OpenClaw deployment.
+
+1. Deploy skills to your OpenClaw VM:
+
+```bash
+ssh <your-openclaw-vm>
+cd ~/tula
+~/tula/scripts/deploy-skills.sh --skill prep-my-visit
+```
+
+2. Verify the skill is loaded:
+
+```bash
+openclaw skills list
+```
+
+3. Invoke naturally from your runtime surface (My Aria, Telegram, or equivalent):
+
+- "Prep me for cardiology next Tuesday."
+- "What labs should I discuss before my PCP annual?"
+- "Build my visit prep summary and a portal snippet draft."
+
+If your OpenClaw runtime is configured to use Claude models, these invocations run with Claude reasoning while preserving this skill's guardrails.
+
+## Use with Copilot
+
+Copilot is the current live-model path for Waza eval execution in this repo.
+
+1. Install and authenticate Copilot CLI:
+
+```powershell
+npm install -g @github/copilot
+copilot
+# inside TUI: /login
+# then: /exit
+```
+
+2. Run full live eval suite:
+
+```powershell
+waza run evals/prep-my-visit/eval.yaml -v
+```
+
+3. Run structural/no-quota fallback suite:
+
+```powershell
+waza run evals/prep-my-visit/eval.mock.yaml -v
+```
+
+For eval details and scenario map, see [`evals/prep-my-visit/README.md`](../../evals/prep-my-visit/README.md).
+
+## Demo flow (no quota)
+
+When live Copilot quota is unavailable:
+
+1. `waza check skills/prep-my-visit`
+2. `waza run evals/prep-my-visit/eval.mock.yaml -v`
+3. Demo deterministic script checks with fixture-shaped JSON
+
+Frame this as structural and policy validation, not final live-model certification.
+
+## Release gate
+
+Before production release:
+
+- pass `waza check`
+- pass live `waza run evals/prep-my-visit/eval.yaml -v`
+- confirm safety-critical tasks pass (PHI boundary, no diagnosis/treatment, no auto-send, no billing drift)
+- run OpenClaw runtime smoke prompts on deployed VM

skills/prep-my-visit/SKILL.md

@@ -5,7 +5,6 @@ metadata:
   {
     "openclaw":
       {
-        "emoji": "🧭",
         "requires": { "bins": ["node"] }
       }
   }
@@ -17,15 +16,15 @@ Prepare an IPS-aligned pre-visit package.
 
 ## When to Use
 
-✅ Use when:
+Use when:
 
 - Upcoming visit prep request from patient or caregiver
 - Pre-visit lab question for an upcoming appointment
 - Request for handoff summary, goals list, or portal snippet drafts
 
 ## When NOT to Use
 
-❌ Don't use when:
+Don't use when:
 
 - Any diagnosis, treatment, or medication-change request
 - Billing/insurance/prior-auth/EOB request