[pswider]

Renames the open-source single-user dashboard from My Aria to My Tula (apps/my-aria/ to apps/my-tula/) and defines three distinct products:

Name	What it is	License
Tula	Personal health agent, its skills, and the My Tula dashboard, single user, self-hosted	Apache-2.0, public
My Aria	The patient's agent and app, issued and governed by a healthcare organization	Commercial
Aria	The platform behind My Aria: Patient Swarm runtime, identity, audit, EHR integration, governance	Commercial

Why: "My Aria" on a free public app weakens the name for the product RealActivity sells. Keeping each name on one product makes the trademark protect the commercial side and keeps the open side clearly free.

Changes

• git mv of the app so history follows. Package name, page title, header wordmark, disclaimer copy, log prefix, sidebar storage key, and screenshot script all say My Tula.
• Screenshots regenerated from the production build with synthetic fixtures (1440x900, dark, full page).
• prep-my-visit PDF wordmark renders "My Tula".
• Tenant cloud-init-template.yaml writes .env.local to apps/my-tula/. deprovision.sh clears both the old and new screenshot directories.
• OPEN_CORE.md, TRADEMARK.md, README.md, docs/aria-commercial-platform.md describe the three-name structure. Trademark policy now names My Tula and My Aria explicitly.
• Draft articles (both status: draft) use My Tula, and the "Three names" section in the portal article explains My Aria and Aria as the commercial side.
• Dated history entries (roadmap, demo video, dashboard plan) note the earlier name, so they stay accurate.

My Tula dashboard

Verification

• npm ci, npm run typecheck, npm run build in apps/my-tula/ pass.
• Production server with TULA_USE_FIXTURES=1: /dashboard, /home-devices, /sdoh/ai-chats return 200 and render "My Tula". The only console error is a missing favicon, also present on main.
• python3 -m py_compile on render_visit_brief.py, bash -n on deprovision.sh.
• ASCII check is clean on every line added. docs/dashboard-build-plan.md already contained non-ASCII dashes and arrows on main; left untouched.
• waza check was not run (Waza is not installed in this environment). The skill change is two strings and comments.

Needs a human

• Live VM: after git pull, the systemd unit my-aria.service still points at apps/my-aria/. Follow "Upgrading a running deployment" in apps/my-tula/README.md (copy .env.local, build, repoint WorkingDirectory, restart). The dashboard on port 3002 stays down until this is done.
• Saved sidebar state resets once in the browser because the storage key changed.
• Trademark clearance for "My Tula", "Aria", and "My Aria" in healthcare software before marketing use.
• Already-published copies of the articles or demo video outside this repo still say My Aria.

_{To show artifacts inline, enable in settings.}

  

[github-actions]

Eval status updated for this PR

Preview of docs/evals.md after this PR merges:

Skill Evaluation Status

Continuous evaluation status for Tula skills. This page is regenerated
automatically by scripts/generate-eval-status.sh on every CI run that
touches skills/ or evals/. Static analysis (compliance, spec
checks, token budgets) is fresh on every run; live eval results come
from manually-published runs in results/.

Powered by Microsoft Waza. Standard:
Patient Agent Eval Standard v0.1.

Skill	Tasks	Mock CI	Compliance	Spec	Tokens	Last live run
epic-note	6	yes	Medium-High	9/9 ✓	821 / 500 ⚠	-
health-records	8	yes	Medium-High	9/9 ✓	1410 / 500 ⚠	-
lookout	9	yes	Medium-High	9/9 ✓	1577 / 500 ⚠	-
med-pdf	8	yes	Medium-High	9/9 ✓	930 / 500 ⚠	-
memory-diff	7	yes	Medium-High	9/9 ✓	1183 / 500 ⚠	-
myhealth-pulse	6	yes	Medium-High	9/9 ✓	1176 / 500 ⚠	-
prep-my-visit	16	yes	Medium-High	9/9 ✓	457 / 500 ✓	-
request-amendment	12	yes	Medium-High	9/9 ✓	990 / 500 ⚠	-
composition	6	yes	-	-	-	-

---

What this measures

• Tasks - count of YAML tasks in evals/<skill>/tasks/ (including golden/).
• Mock CI - yes when eval.mock.yaml exists (structural gate on every PR).
• Compliance - Waza's agentskills.io readiness score
  (High / Medium-High / Medium / Low). Medium-High or better
  is the house target.
• Spec - count of agentskills.io spec checks the skill passes
  (spec-frontmatter, spec-name, spec-allowed-fields, and so on).
  9/9 is full pass.
• Tokens - total tokens in SKILL.md against Waza's 500-token soft
  limit. Tula's house style accepts a higher count when openclaw
  fidelity would suffer (per skills/AGENTS.md's "Token Discipline"
  section). ⚠ marks "exceeds the soft cap but intentional"; ✓ marks
  "within budget."
• Last live run - most recent waza run output published in
  results/. Cells show pass rate, run date, and model used (e.g.,
  5/5 ✓ (2026-05-17, sonnet-4.6)). Live eval execution requires
  executor: copilot-sdk plus model auth, so it is a deliberate
  publish today rather than a per-PR CI run. Raw run outputs stay
  private; only the pass-rate summary surfaces here.

What this does NOT measure

• The model's actual answer quality. Evals check task-completion
  signals (output shape, presence/absence of keywords, routing
  behavior, schema validity), not clinical correctness.
• Production behavior under PHI. All evals run against synthetic
  personas. See evals/*/fixtures/ for the test data.
• Anything inside Aria's closed governance layer - multi-tenant
  isolation, audit emission, cross-actor coordination - which is
  evaluated separately under hospital-scale fixtures.

See also

• Patient Agent Eval Standard v0.1
• Eval suites - task definitions and fixtures
• Skill authoring conventions
• Tula deployment guide
• Microsoft Waza - the eval framework

[github-actions]

Eval status updated for this PR

Preview of docs/evals.md after this PR merges:

Skill Evaluation Status

Continuous evaluation status for Tula skills. This page is regenerated
automatically by scripts/generate-eval-status.sh on every CI run that
touches skills/ or evals/. Static analysis (compliance, spec
checks, token budgets) is fresh on every run; live eval results come
from manually-published runs in results/.

Powered by Microsoft Waza. Standard:
Patient Agent Eval Standard v0.1.

Skill	Tasks	Mock CI	Compliance	Spec	Tokens	Last live run
epic-note	6	yes	Medium-High	9/9 ✓	821 / 500 ⚠	-
health-records	8	yes	Medium-High	9/9 ✓	1410 / 500 ⚠	-
lookout	9	yes	Medium-High	9/9 ✓	1577 / 500 ⚠	-
med-pdf	8	yes	Medium-High	9/9 ✓	930 / 500 ⚠	-
memory-diff	7	yes	Medium-High	9/9 ✓	1183 / 500 ⚠	-
myhealth-pulse	6	yes	Medium-High	9/9 ✓	1176 / 500 ⚠	-
prep-my-visit	16	yes	Medium-High	9/9 ✓	457 / 500 ✓	-
request-amendment	12	yes	Medium-High	9/9 ✓	990 / 500 ⚠	-
composition	6	yes	-	-	-	-

---

What this measures

• Tasks - count of YAML tasks in evals/<skill>/tasks/ (including golden/).
• Mock CI - yes when eval.mock.yaml exists (structural gate on every PR).
• Compliance - Waza's agentskills.io readiness score
  (High / Medium-High / Medium / Low). Medium-High or better
  is the house target.
• Spec - count of agentskills.io spec checks the skill passes
  (spec-frontmatter, spec-name, spec-allowed-fields, and so on).
  9/9 is full pass.
• Tokens - total tokens in SKILL.md against Waza's 500-token soft
  limit. Tula's house style accepts a higher count when openclaw
  fidelity would suffer (per skills/AGENTS.md's "Token Discipline"
  section). ⚠ marks "exceeds the soft cap but intentional"; ✓ marks
  "within budget."
• Last live run - most recent waza run output published in
  results/. Cells show pass rate, run date, and model used (e.g.,
  5/5 ✓ (2026-05-17, sonnet-4.6)). Live eval execution requires
  executor: copilot-sdk plus model auth, so it is a deliberate
  publish today rather than a per-PR CI run. Raw run outputs stay
  private; only the pass-rate summary surfaces here.

What this does NOT measure

• The model's actual answer quality. Evals check task-completion
  signals (output shape, presence/absence of keywords, routing
  behavior, schema validity), not clinical correctness.
• Production behavior under PHI. All evals run against synthetic
  personas. See evals/*/fixtures/ for the test data.
• Anything inside Aria's closed governance layer - multi-tenant
  isolation, audit emission, cross-actor coordination - which is
  evaluated separately under hospital-scale fixtures.

See also

• Patient Agent Eval Standard v0.1
• Eval suites - task definitions and fixtures
• Skill authoring conventions
• Tula deployment guide
• Microsoft Waza - the eval framework