Template's PhotometryFilters.json ships Paranal/VISTA bands with no matching Paranal row in Telescopes.json/Instruments.json

**Where:** The astrodb-template-db template's shipped reference data: `data/reference/PhotometryFilters.json` (inherited as-is from the template, never edited by this project) already contains `Paranal/VISTA.Y`, `Paranal/VISTA.J`, `Paranal/VISTA.H`, `Paranal/VISTA.Ks` (and other pre-populated bands like `PAN-STARRS/PS1.y`, `UKIRT/WFCAM.*`, `UKIRT/UKIDSS.*`, `UKIRT/UFTI.*`), but `data/reference/Telescopes.json` has no `Paranal` row at all (only `VLT-UT1..4` for that facility) and `Instruments.json` has no matching `Paranal` instrument row either.

**What happened:** `astrodb-ingest-photometry`'s Step 4 pattern (skip `ingest_photometry_filter` if the band already exists in `PhotometryFilters`, since `ingest_photometry_filter` raises "already exists") means the Telescopes/Instruments setup that `ingest_photometry_filter` would normally do as a side effect never ran for these pre-shipped bands. `ingest_photometry(..., telescope="Paranal", ...)` then failed for every VHS/VIKING/VVV/VIRAC2 magnitude with `Telescope Paranal not found in Telescopes table.` — 299 rows skipped on the first dry run.

**Workaround:** Added an explicit `ingest_instrument(db, telescope=telescope, instrument=instrument, mode="Imaging", raise_error=False)` call for every band in `BAND_SETUP` *before* the "does the filter already exist" check, regardless of whether `PhotometryFilters` already has the band. `ingest_instrument` is idempotent, so this is cheap and safe to always run.

**Suggested change:** Either (a) fix the shipped `data/reference/Telescopes.json`/`Instruments.json` in the upstream `astrodb-template-db` template to include `Paranal`/`VISTA` (and audit for other pre-populated `PhotometryFilters` bands with the same gap), or (b) update `astrodb-ingest-photometry`'s Step 4 instructions to call `ingest_instrument` unconditionally for every band's `(telescope, instrument)` pair before the "band already exists, skip" check, rather than only relying on `ingest_photometry_filter`'s side effect (which is skipped precisely when the band already exists).

---
_Reported from a gotchas.md log filed by a skill user (2026-08-28)._

Compare data/reference/PhotometryFilters.json with data/reference/Telescopes.json and data/reference/Instruments.json, then read astrodb-ingest-photometry's Step 4 and the BAND_SETUP workaround described here. Determine whether the fix belongs in the upstream template data or the Step 4 instructions; done means Paranal/VISTA bands have matching setup and the reported photometry rows are no longer skipped.

Written by the indexing model from the issue text.