fix(android): publish covered state from exact order evidence

[thymikee]

Summary

Centralize Android covered-state publication without guessing z-order.

• Keep nodes in regular snapshots and make the daemon the sole owner of interactionBlocked: "covered" publication.
• Carry API 24+ sibling drawing order through one private, non-serializable evidence sidecar. The evidence includes the original sibling group, so presentation wrapper collapse cannot compare unrelated nodes.
• Fail conservative on API 23, where Android does not expose sibling drawing order, and retain androidSnapshot.occlusionScanUnavailable: true rather than claiming API-level neutrality.
• Replace recursive and semantic-order heuristics with a bounded iterative classifier. Sparse overlays remain actionable, 5,000-node deep trees cannot overflow the stack, and exhausted evidence publishes no covered verdict.

This advances #1832 but does not close it: API 23 cannot provide the exact order needed for a reliable covered-state claim.

Validation

• Planted red/green: sparse floating overlay was falsely classified as a replacement surface without exact order.
• Planted red/green: presentation-reparented nodes from different source sibling groups were incorrectly compared.
• Planted red/green: a 5,000-node deep capture overflowed the recursive footprint builder.
• Focused contracts, Android capture, daemon publication, and classifier suite: 114 tests passed.
• pnpm check:affected --run: all runnable checks passed, including format, lint, typecheck, layering, fallow complexity/dead-code audit, build, command-doc coverage, and 547 related test files / 4,446 tests.

[github-actions]

Size Report

Metric	Base	Current	Diff
JS raw	2.37 MB	2.38 MB	+812 B
JS gzip	797.9 kB	798.2 kB	+316 B
npm tarball	919.3 kB	919.5 kB	+248 B
npm unpacked	3.19 MB	3.19 MB	+804 B

npm unpacked components

Component	Base	Current	Diff
JS / dist source	2.52 MB	2.52 MB	+804 B
Apple runner source/project	564.3 kB	564.3 kB	0 B
macOS helper source	54.5 kB	54.5 kB	0 B
Android helper artifacts	0 B	0 B	0 B
Other package files	44.5 kB	44.5 kB	0 B

Startup median (7 runs, lower is better):

Scenario	Base	Current	Diff
CLI --version	29.3 ms	29.1 ms	-0.2 ms
CLI --help	79.9 ms	79.8 ms	-0.1 ms

Top changed chunks:

Chunk	Raw diff	Gzip diff
dist/src/session-snapshot.js	+2.4 kB	+774 B
dist/src/snapshot.js	-2.0 kB	-601 B
dist/src/runtime.js	+4 B	+7 B
dist/src/session2.js	+4 B	+6 B
dist/src/internal/daemon.js	0 B	+2 B

Top changed packed files

Packed file	Base	Current	Diff
dist/src/session-snapshot.js	34.1 kB	36.6 kB	+2.4 kB
dist/src/snapshot.js	31.3 kB	29.2 kB	-2.0 kB
dist/src/snapshot-private-evidence.js	0 B	425 B	+425 B
dist/src/snapshot-clickability.js	164 B	0 B	-164 B
dist/src/snapshot-chrome.js	7.9 kB	8.0 kB	+124 B
dist/src/sdk-selectors.d.ts	26.0 kB	26.0 kB	-8 B
dist/src/android.js	16.7 kB	16.7 kB	+4 B
dist/src/device-claim-conflict.js	87.8 kB	87.8 kB	+4 B
dist/src/runtime.js	64.2 kB	64.2 kB	+4 B
dist/src/session2.js	230.1 kB	230.1 kB	+4 B

[github-actions]

PR Preview Action v1.8.1
🚀 View preview at https://callstack.github.io/agent-device/pr-preview/pr-1981/
Built to branch gh-pages at 2026-08-23 17:27 UTC. Preview will be ready when the GitHub Pages deployment is complete.

[thymikee]

Not ready at 91281d2c:

1. The migration drops the old classifier’s multi-rect union semantics. android-replacement-surface-occlusion.ts reduces a covering subtree to one strongestPaint rect, whereas the prior owner measured the union of every paint rect. Two higher-order actionable descendants can collectively cover a stale surface >=90% while each alone covers <90%, leaving the stale action hittable. Restore bounded union coverage and add a planted regression for collective descendant coverage.
2. Coverage is deterministically red: the test-file-size ratchet reports growth in src/platforms/android/__tests__/snapshot.test.ts (1495→1542) and src/daemon/handlers/__tests__/session-replay-divergence.test.ts (1215→1218). Move the new scoped evidence/publication case to a focused sibling test owner and avoid growing the replay aggregation.
3. Before readiness, attach the Android native-a11y → snapshot: conform to the acquire/present engine contracts (sibling of #1797) #1832-required 33-screen true-positive corpus/consumer inventory and live Pixel evidence for this changed path. Current fixture evidence is insufficient for removing the platform pruner.

The private evidence route and size impact otherwise look sound.