feat(github-app): gate false-positive telemetry

[JSONbored]

Parent roadmap: #525
Parent phase: #528

Background

Maintainers won't move gates from advisory to block without evidence the gates are precise. We need to measure false positives.

Goal

Gate false-positive telemetry: record gate-blocked-then-merged (override) events and expose a per-gate-mode false-positive rate.

Current Behavior

No gate-outcome telemetry; overrides aren't correlated with eventual merges.

Desired Behavior

• Record when a gate-blocked PR is later merged (or overridden via feat(github-app): gate appeal/override command + configurable settings #538), and aggregate a false-positive rate per gate type. Reuses recordProductUsageEvent + the recommendation-outcomes service.

Implementation Requirements

• New telemetry table/migration; aggregation service. No PII.

Public/Private Output Boundaries

• Internal/maintainer-authenticated; no public exposure; no reward/trust fields.

Acceptance Criteria

• A blocked-then-merged PR increments the false-positive count; rate is queryable per gate mode.

Testing Requirements

• npm run test:ci, 97%+ coverage. Telemetry-privacy + aggregation tests.