Design: telemetry story for v5

Spec the telemetry story for v5 from scratch, independent of what
exists in vnext today. Define what we want to measure, the vendor /
pipeline, opt-in/opt-out UX, privacy guarantees, and the event
taxonomy. A follow-up issue will audit the current vnext
implementation against this spec and call out gaps.

## Definition of Done

- [ ] Goals + non-goals documented (what telemetry is for, what it
is not)
- [ ] Vendor + pipeline decision recorded (OTel exporter,
destination, retention)
- [ ] Opt-in/opt-out model defined (env var, settings toggle,
first-run UX)
- [ ] Event taxonomy: per-command timing, outcomes, errors, workload
events, version/runtime fingerprint
- [ ] PII / secret scrubbing rules specified
- [ ] Sampling + rate-limiting rules specified
- [ ] Follow-up issue opened to audit vnext against this spec

No files, tests, or entry points are named. Start from the issue checklist and structure a standalone v5 telemetry specification covering goals, vendor and pipeline, consent UX, event taxonomy, privacy, sampling, and rate limits. Done means every checklist item is documented and a follow-up issue is opened to audit the current vnext implementation.

Written by the indexing model from the issue text.