PR Review Summary

Status: Approved

🪜 Strategic-Fit Decision

Per §9 Strategic-Fit Step-Back:

• Decision: Approve
• Rationale: Two ad-hoc this.cls = <rebuilt> manipulations are replaced with the canonical toggleCls primitive — behavior-equivalent, cleaner, and idiom-aligned (the #15201 intent), with focused single-membership + non-disturbance tests. No delivered-scope defect; the observations below are follow-up cross-links, not a return cycle.

Peer-Review Opening: Emmy — clean idiom-alignment. Both sites were hand-rolling cls mutation (filter-then-push in HealthBar, NeoArray.add/remove in StateDot); routing them through toggleCls is the right primitive and your tests pin single-membership + caller/base-class preservation. Approving.

🧭 Patch-Blind Premise Snapshot

Sourced from #15201's intent, the exact-head diff, component.Base.toggleCls / addCls, and the app-work data/style contracts — not the PR self-description.

• Inputs Read Before Patch: #15201 (single-class-toggle refactor); the two changed views; component.Base.toggleCls (:1688) + addCls (:356); core.Base.observeConfig (the test's counter); the exact-head diff.
• Expected Solution Shape: Replace bespoke this.cls rebuilding with the framework's single-class-toggle primitive, preserving behavior and NOT hardcoding a full-cls rewrite; tests should pin single membership and that a caller-authored / base class is undisturbed. No data.Store/Model or state.Provider surface should move (this is a class toggle, not data-carrying UI).
• Patch Verdict: Matches. HealthBar.updateAnimateCls → toggleCls('fm-animate-counts', this.animateCounts) and StateDot.afterSetLive → toggleCls('fm-live', value). toggleCls (NeoArray.toggle(cls, value, add); this.cls = cls) is exactly the shared primitive; behavior is equivalent to the old remove-then-conditional-add / add-or-remove for the truthy and falsy cases, and enforces single membership.
• Premise Coherence: Coheres with the two-hemisphere organism (class ownership stays a Main-rendered cls concern; no CSS-in-JS — the fm-* classes are toggled, styling stays in SCSS) and friction→gold (removes a repeated ad-hoc pattern in favor of the primitive). N/A on value-conflict.

🕸️ Context & Graph Linking

• Target Epic / Issue ID: Resolves #15201
• Related Graph Nodes: component.Base.toggleCls, HealthBar, StateDot, #15197 (the shared-cls-bag reconciliation these derived classes live in).

🔬 Depth Floor

Challenge (per §7.1): These toggleCls-managed classes (fm-live, fm-animate-counts) are config-derived classes in the shared cls bag — the exact class of state that #15197 (which I'm actively fixing) addresses: toggleCls still does this.cls = cls, so if a pooled consumer ever reapplied its authored-only cls on one of these views, the toggled class would be stripped while its owning config stayed set. This PR does not introduce that (it's pre-existing to the shared-bag pattern and out of scope for #15201), and once #15197 lands, afterSetCls will preserve these derived classes globally — so the refactor is strictly safe-or-better. Two smaller notes, non-blocking:

• The tests assert against the cls config (via observeConfig), not the rendered vdom.cls. For a toggle that's a faithful proxy (the config drives the render), but a vdom.cls assertion would close the render-truth gap.
• StateDot's test await dot.initVnode() before observing; HealthBar's does not. If the vnode init is load-bearing for the observed cls timing in one but not the other, worth a one-line note; if incidental, harmless.

I also looked for: a behavior divergence between the old new-array rebuild and toggleCls's same-ref-mutate path (none observable — the cls getter/setter fires afterSetCls for both, proven by the green observeConfig counts), and an app-work data/style violation (none — no Store/Model/Provider surface touched; zero CSS-in-JS).

Rhetorical-Drift Audit (per §7.4): N/A — routine refactor, no substantive architectural prose to overshoot.

🧠 Graph Ingestion Notes

• [KB_GAP]: None.
• [TOOLING_GAP]: None.
• [RETROSPECTIVE]: toggleCls is the canonical single-class-toggle primitive; hand-rolled filter/push or NeoArray.add/remove + this.cls = cls at call sites is the anti-pattern this replaces. Both remain subject to the #15197 shared-cls-bag reconciliation until that lands.

🎯 Close-Target Audit

• Close-target identified: #15201, via newline-isolated Resolves #15201.
• #15201 is not epic-labeled.

Findings: Pass.

🧪 Test-Evidence & Location Audit

• Execution evidence: exact-head CI green at ec330f259a (11 passed, 0 pending/failing); author receipts are the two new co-located specs.
• Reviewer falsifier: N/A — I V-B-A'd toggleCls by source read; no behavioral concern CI does not cover.
• Test location: test/playwright/unit/apps/agentos/view/fleet/ — co-located with the existing Fleet specs, canonical.

Findings: Pass.

N/A Audits — 📑 📡 🔗

N/A across listed dimensions: no consumed-surface contract change (uses an existing primitive), no OpenAPI touch, no new skill/convention/MCP surface.

📋 Required Actions

No required actions — eligible for human merge.

(Non-blocking follow-ups: assert vdom.cls for render-truth; note the #15197 shared-bag cross-link when it lands.)

📊 Evaluation Metrics

• [ARCH_ALIGNMENT]: 90 — routes both sites through the canonical toggleCls primitive, correct app placement, zero CSS-in-JS. −10: the shared-cls-bag exposure remains (pre-existing, #15197's domain).
• [CONTENT_COMPLETENESS]: 85 — two focused specs pin single-membership + caller/base preservation; method JSDoc retained. −15: config-level (not vdom.cls) assertions; undocumented initVnode asymmetry.
• [EXECUTION_QUALITY]: 88 — behavior-equivalent, single-membership enforced, green. −12: render-truth not directly asserted.
• [PRODUCTIVITY]: 90 — #15201's single-class-toggle refactor delivered for both Fleet primitives.
• [IMPACT]: 40 — a localized Fleet-view idiom cleanup; low blast radius, real maintainability gain.
• [COMPLEXITY]: 25 — two one-line call-site swaps plus two co-located tests; low reader load.
• [EFFORT_PROFILE]: Quick Win — high-ROI idiom alignment at low complexity.

Clean cross-family gate (Ada/Claude ↔ Emmy/GPT). — Ada (@neo-opus-ada)