Skip inserter-hidden Jetpack blocks in catch-all discovery

[lezama]

Related issues

• Follows up on review feedback from Add Jetpack catch-all rule for non-core feature requests #3362: Add Jetpack catch-all rule for non-core feature requests #3362 (comment)

How AI was used in this PR

Claude (Opus 4.7) drafted the change and this description. I steered the wording — the comment "kept registered for back-compat" was too narrow, since supports.inserter: false is also used for parent-only blocks, programmatic-insert-only blocks, internal blocks, etc. The honest framing is just "the user can't insert these from the editor, so they aren't candidates."

Proposed Changes

The Jetpack catch-all rule's discovery one-liner enumerates every jetpack/* block in the registry. But registered ≠ available in the editor: Jetpack (and other plugins) keep some blocks registered with supports.inserter: false because they're parent-only, programmatic, or deprecated. Recommending those blocks would have the agent generate markup the user can't reproduce in the editor.

This PR adds an isset/!== false guard around $b->supports["inserter"] so hidden blocks drop out of the listing, and updates the surrounding sentence to explain why.

Testing Instructions

• npm run typecheck — passes locally.
• The existing eval covers this rule: eval/promptfoo.config.yaml → jetpack-catchall-slideshow. It asserts the agent uses any jetpack/* block on a slideshow request, which is unaffected by the filter (the slideshow block is inserter-visible).

Pre-merge Checklist

• Have you checked for TypeScript, React or other console errors?

[wpmobilebot]

📊 Performance Test Results

Comparing b248647 vs trunk

app-size

Metric	trunk	b248647	Diff	Change
App Size (Mac)	1408.66 MB	1408.66 MB	+0.00 MB	⚪ 0.0%

site-editor

Metric	trunk	b248647	Diff	Change
load	1495 ms	1498 ms	+3 ms	⚪ 0.0%

site-startup

Metric	trunk	b248647	Diff	Change
siteCreation	8081 ms	8098 ms	+17 ms	⚪ 0.0%
siteStartup	4936 ms	4931 ms	5 ms	⚪ 0.0%

---

Results are median values from multiple test runs.

Legend: 🟢 Improvement (faster) | 🔴 Regression (slower) | ⚪ No change (<50ms diff)