Upgrade WeChat automation and incoming call decisions

[bjfwan]

Summary

• Upgrade WeChat video automation with semantic UI snapshots, task-engine decisions, scored contact matching, and failure replay persistence.
• Add incoming auto-answer compatibility decision engine and wire it into the phone-state receiver.
• Improve elder-friendly phone/video contact actions with larger buttons, clearer phone/video colors, accessible labels, and refreshed release APK/docs for v1.9.0.
• Expand tests from single-function coverage into logic-chain coverage for WeChat tasks, failure replay, incoming call decisions, and action button UI.

Verification

• ./build.bat :app:testDebugUnitTest ✅ 309 tests, 0 failures
• ./build.bat :app:lintDebug ✅
• ./build.bat :app:assembleRelease ✅
• docs/app-release.apk SHA-256: 29072C0BC5F951AB369FE282BC27DB14E53E8CA08EDF27AA1104B36FA8143718

Release

• Version: 1.9.0
• Version code: 16
• APK size: 2.16 MB

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d0e85a05a0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Stop penalizing exact contact hits for network section marker

Subtracting 50 points when displayName == normalizedName and the page contains "搜索网络结果" causes valid local contact matches to be rejected in common WeChat result layouts that show both local hits and a network-results section on the same screen. In that scenario the score drops from 80 (title_id + exact_title + clickable) to 30, so accepted becomes false and the automation can fail with no_contact_result even though the target contact is present.

Useful? React with 👍 / 👎.