feat: add light-link-studio example witnessing receiver collections in pixels

[TMHSDigital]

Summary

The light-linking collection witness from the ROADMAP pool: one key linked to a hero sphere beside an identical decoy, with the entire contract proven in pixels by two renders in a single check pass.

Contract witnessed

1. The API is on the light OBJECT — ld.light_linking is an AttributeError on both versions (verified); the path is obj.light_linking.receiver_collection (ObjectLightLinking, with blocker_collection alongside). Assignment reads back.
2. Linked, the key lights only the receiver collection — hero/decoy luminance ratio 4.0x (gate ≥ 3x), sampled at each sphere's projected center so framing can't move the sample off the subject (a framing iteration broke exactly that during authoring).
3. Unlinked in the same check, the restriction vanishes surgically — decoy rises 244% (5.1.2) / 251% (4.5.11) while hero drift is 0.0%. The shipped check demonstrates both states, not just one.
4. Verified engine note — light linking also works on EEVEE Next: 5.5x linked ratio on 4.5.11 EEVEE Next, 5.9x on 5.1.2 EEVEE, matching Cycles within sampling noise. The "EEVEE ignores it" assumption is false; the check still pins Cycles for deterministic tiny-sample CPU renders, and view_transform = 'Standard' for meaningful luminance ratios.

Falsification evidence (temporary breaks, restored after measurement):

Probe	Break	Exit	Measured
A	link inverted to the decoy collection	6	hero drops to 0.179, ratio collapses to 0.22x
B	EEVEE engine substitution	0 (documented, not a failure)	linking honored on EEVEE Next on both versions (5.5x / 5.9x)

Contact-sheet gate: light-link-studio-contact-sheet.webp — verdicts: stage darkness ✓, wedge warmth ✓, subject fill ✓, saturation ✓ (blazing orange hero), thumbnail legibility ✓✓ (strongest contrast in the lineup). Mean luminance 111.2 vs calibration 139.4 / 100.1 / 143.4 — inside range.

Proven by live run vs inspection:

• Live run, local: check exit 0 on .scratch/blender-5.1.2-windows-x64/blender.exe (Blender 5.1.2, hash ec6e62d40fa9) and .scratch/blender-4.5.11-windows-x64/blender.exe (Blender 4.5.11 LTS, hash 4db51e9d1e1e) — matching values (ratio 4.0x both; rise 244% vs 251%; drift 0.0%). Render exit 0 on both binaries (Cycles), pixels iterated (5 drafts: wedge flooding the decoy, sample-point fragility fixed by projected-center sampling). All probes live.
• Live run, CI: both smoke series on the PR head SHA — see checks.
• Inspection only: nothing claims otherwise.

Test plan

• Validate, validate-manifest, validate-counts (README substring 30 examples)
• Blender 5.1 smoke + Blender 4.5 smoke — new step runs examples/light-link-studio/light_link_studio.py --
• Ecosystem drift; Socket Security
• Local binaries: .scratch/blender-5.1.2-windows-x64/blender.exe (5.1.2), .scratch/blender-4.5.11-windows-x64/blender.exe (4.5.11 LTS)

Integration: example dir + README, gallery.json entry, plugin manifest, smoke step, top-level README ("Mesh, curves & text" category now 10, counts 29→30), generated pages read character-by-character (alt and callout full), ROADMAP entry shipped-marked.

🤖 Generated with Kimi Code