[prakashUXtech]

This is the last PR of #436 (retouch-heal, slice RH-z). Closes #436.

What

docs/blueprints/core/features/retouch-heal.md (RH-8): the status line now says built.

It lists the merged slice PRs:

• feat(coreraster): declare the retouch-heal contract stubs #560, test(raster): retouch-heal acceptance suites, disabled until RH-3 to RH-6 #591, test(retouch): fixtures for the retouch-heal checks #566 (RH-a)
• feat(raster): heal through a mask with a gradient-domain solve #601, feat(raster): fill from surroundings by seeded patch search and heal #613 (RH-heal)
• feat(raster): dodge, burn and sponge tone dabs #606, feat(raster): smudge dabs drag a carried colour that fades by strength #602 (RH-dab)
• test(raster): retouch-heal walk and budgets #616 (RH-z)

It also lists the review fixes: #644, #657, #659, #661, #677, #684, #687, #710, #711, #712, #769.

It records where the code differs from the packet:

• The heal solves with a multigrid V-cycle. It stops when a whole cycle moves no cell by the tolerance.
• The fill heals the source found by a per-pixel patch search inside a search window.
• The release budgets run and report, and a miss is recorded as a known issue. The 200 px heal dab took 16.7 to 19.3 ms against 12 ms on a loaded Linux host (perf(heal): heal dab over its 12 ms release budget; share one CoreRaster SplitMix64 #770). The Mac finisher times it.
• The packet's FillRandom follow-up is now issue perf(heal): heal dab over its 12 ms release budget; share one CoreRaster SplitMix64 #770.

Why

RH-z closes the feature: the walk, the budgets and the round-4 fixes (#769) are already on dev, and this PR marks the packet built.

Tests

This PR changes docs only. scripts/ci printed CI: PASS on this head, on dev 3366440. The packet's check for RH-8 is "status line says built".

Belt-Lead: act-1a157fd0281-warg
Feature: retouch-heal
Slice: RH-z
Mutate: 0/0 none (docs only)
Fast-Check: pass
Head: 7b397a6

[github-actions]

Fast lane: merged into dev (11 lines, 1 files, Fast-Check marker).