[Story]: GitHub Copilot Metrics API connector for adoption and aggregate acceptance

## Summary
Build a scheduled connector pulling GitHub Copilot's org-level Metrics API (`/orgs/{org}/copilot/metrics`) for adoption and aggregate acceptance: `total_suggestions_count`, `total_acceptances_count`, `total_active_users` for `copilot_ide_code_completions`, sliced by editor/language/model.

## Intent / Source of truth
GitHub's documented Copilot Metrics API. 2026-08-13 backlog consolidation, sales-accuracy constraint identified during vendor capability review.

## Scope
- [ ] Scheduled pull of `/orgs/{org}/copilot/metrics`
- [ ] Ingest `total_suggestions_count`, `total_acceptances_count`, `total_active_users` sliced by editor/language/model for `copilot_ide_code_completions`
- [ ] Map into the schema from story 1, tagged clearly as aggregate-only (no per-user field populated)
- [ ] Document the 5-seat-per-day minimum constraint in the connector's own README/docstring so it can't be silently forgotten downstream

## Out of scope
- Any attempt to infer per-individual-user acceptance from aggregate data — that is not possible and must not be implied
- Copilot Chat or Copilot Business features beyond `copilot_ide_code_completions`

## Verification
Live pull against a real org showing non-empty metrics once the 5-seat minimum is met; a documented, explicit statement (in the dashboard UI and any sales collateral) that this connector is aggregate-only.

## Risk assessment
CONSTRAINT to state plainly and enforce in code/docs: this is aggregate only, never per-individual-user, and GitHub enforces a 5-seat-per-day minimum before returning any data. Do not promise per-user Copilot acceptance in a sales deck — this has been flagged explicitly to prevent that overclaim.

## AI Usage Declaration
Drafted with AI assistance during the 2026-08-13 cross-repo backlog consolidation. A human owns intent, verification and consequences; acceptance criteria must be verified against the live system before this is closed.