whey Claude telemetry has never been collected: argparse default points at an absent sqlite

collect.py:694 sets --claude-db default to ~/.local/share/token-usage/token-usage.sqlite, which does not exist on whey. agent-telemetry-collect.service (user scope, every 15 min) passes no --claude-db, so the open raises sqlite3.OperationalError before any table inspection and lands in the skipped_absent branch at :723-730, which hardcodes source_kind='claude_sqlite_hourly'. The daily-vs-hourly schema branch at :227-233 is therefore unreachable on this host. The file's own header at :117-122 already documents this exact path confusion. whey/claude has never been collected. DO NOT simply repoint --claude-db: whey's real store ~/.local/share/claude-usage/claudeUsage.sqlite (tables dailyUsage/usagePolls/usageAlerts, no usage_hourly) has mtime 2026-06-06 and is frozen 3 months, so repointing would backfill a dead file and manufacture coverage without data. The live source must be identified first. SECOND, RELATED ITEM FOR THIS WI: one-db expected_coverage declares whey/claude zero rows as expected, so the misconfiguration reads as intended behaviour to every audit. A coverage row must record WHY a zero is expected - no source present vs source present but idle - or an absent collector and a quiet one stay indistinguishable. Binding on the SAVETOKE-33 side (pm-llmmsgsrv-cc): whey agents render as NO COVERAGE, never as a zero or a good score.