docs(plans): validate external wing/nacelle proposal; federate into nav-wire and power-routing plans

[Stab-Rabbit-coding]

Summary

• Validates an external (Gemini) conversation proposing wing/nacelle changes — a "16 mm Unified Spar" with an 11 mm bore for 4× 10 AWG ESC power leads, a static AS5600 Hall encoder on the wingtip, and a removable wingtip "garage" service hatch — against this repo's own analysis and as-built geometry.
• AS5600 and the wingtip hatch are rejected on solid grounds: AS5600 was already datasheet-rejected (on-axis-only, incompatible with the off-axis through-shaft mechanism — the selected part is AK7455); no wingtip hatch exists, and one would conflict with the already-adopted nacelle-off-spar service architecture.
• The power-wire claims turned out to be correct, and my first pass here got that wrong. I initially concluded the wing's existing Ø7 mm double-D conduit was "already a dedicated, sized, working path" for the 4× 10 AWG ESC feeds — that never checked real wire OD against bore area. Verified properly (and independently corroborated by docs/plans/2026-08-27-nacelle-wiring-plan.md, the repo owner's own draft, found merged onto this branch mid-review): even 2 wires at any realistic 10 AWG silicone OD (4.5–5.5 mm) cannot fit one Ø7 mm bore (need ≥2× OD ≈ 9–11 mm at zero clearance). Separately, the nacelle's harness_exit_port() is a rigid slot offset 17.65 mm from the tilt pivot, sweeping a ~29–45 mm arc across the tilt range with no service loop or disconnect modeled anywhere. Both are real, open problems — and the owner's own proposed fix (16 mm OD / 11 mm bore, on-axis) is itself undersized against the same packing math (4 wires need ≥13.3 mm at zero clearance).
• The nav-light wiring finding from the first pass stands: nav_wire_channel() still routes the WS2812C signal wire into the ESC/harness bundle at harness_exit_port() (built 2026-07-04), never updated to match the later (2026-07-18) decision to route it through the spar's hollow bore instead.

Result — three federated plan docs, split by concern:

1. docs/plans/2026-08-29-001-nacelle-nav-wire-hall-reconciliation-plan.md — corrected and narrowed to its still-valid scope: nav-light wire routing + finishing the AK7455 Hall-sensor wing-side geometry (unaffected by how power routing resolves).
2. docs/plans/2026-08-29-002-nacelle-esc-power-routing-plan.md — new. Decision-support plan (same posture as docs/TILT_SPAR_ANALYSIS.md) laying out three candidate architectures for the confirmed power-wire/tilt-articulation problem: (A) keep the torque spar, fix the wire path with a real service loop; (B) enlarge the spar to carry torque+wire on-axis (re-derived, and still undersized even corrected, plus reopens the CF-keyway/bearing/duct-blockage gates TILT_SPAR_ANALYSIS.md already closed); (C) decouple torque transfer from wire routing via a belt/sprocket drive (the owner's stated preference, and the only option that resolves the sizing conflict rather than trading it for a different one). Gated on pinning real wire product data and an owner decision (U4) before any SCAD geometry changes. Also formally closes the long-open U6 item in docs/plans/2026-08-26-001-nacelle-esc-intake-integration-plan.md.
3. docs/plans/2026-08-27-nacelle-wiring-plan.md — the owner's original draft, annotated in place (not rewritten/deleted) pointing to ...002 as the plan of record, per the repo's existing supersession convention.

This PR is docs-only; no SCAD/geometry changes are made here — those are gated behind the owner decision in ...002's U4.

Test plan

• N/A — documentation-only change (no geometry/code touched)
• Follow-up work: pin real WIRE-10AWG product OD/flex rating (...002 U1), owner decision among Options A/B/C (U4), then implement + run the listed verification-contract gates in whichever plan(s) apply