feat(ground): complete regional execution result readback

[d-sorgcodexagent]

What changed

• expand matched PyQt6 and React readback of strict Python-produced regional execution evidence
• present carry, bounce-air, skid, roll, surface-path, total, final downrange/offline, bounce count, ground time, completion, surface transitions, provider/model/calibration identity, ordered observed phases, typed warnings, provenance, and qualification limits
• retain incomplete-run endpoint censorship and label partial results as observed endpoints
• preserve null-result cancellation/failure as unavailable instead of fabricating metrics
• update the campaign manifest, SPEC, and canonical handoffs in the implementation commit

Why

The initial evidence panel established strict plan-bound import but exposed only a subset of the qualified execution envelope. This slice makes the frozen result useful for engineering review while preserving scientific limitations and keeping both clients presentation-only.

Validation

• implementation review: 208 Python tests and 111 React files / 691 tests passed
• independent root review: 5 Python and 5 React focused tests passed
• Ruff/format, strict MyPy, TypeScript, zero-warning ESLint, production build, manifest, documentation, module/changed-file budgets, placeholder scan, and diff checks passed

Remaining work

This remains import-only. UI input/bounce-prefix construction, executor invocation, trajectory/event tables, export/playback, measured calibration, compiled parity, downstream parity, protected CI/review, and release gates remain open. This does not complete #4271 or #4267.

Addresses #4271
Part of #4267

[dieterolson]

Consolidated into #4409. Closing to reduce CI/CD runner load.