N4: Mobile Search retained Cloud messages with exact-message focus #2197

Parent

#2193

What to build

Extend the existing Search journey to authorized Cloud cache content, including channel-only guests, and open/highlight the exact matched message through the real cloud channel screen. Verify mixed All results stay isolated by service/account/Organization/Channel.

Acceptance criteria

• Organization scopes search only their authorized retained messages; All groups Computer and Cloud matches with distinguishable source labels and truthful coverage.
• Cloud cache catalogue is owned/scoped, not a raw scan exposing signed-out files. Account/config changes, logout and explicit permission denial invalidate results and late responses.
• Channel-only guests can search/open only granted Channels; Organization tab presence never grants organization-wide data.
• Add a source-bound Cloud message locator and actual scroll/highlight behavior; result opens the exact message even with colliding IDs in another service/account.
• Eviction or unavailable local target produces an explicit missing-target result; explicit online recovery reauthorizes source access. No silent newest-message or wrong-Channel fallback.
• Computer Search remains functional; query/scope/scroll return and fresh-entry reset rules remain identical across sources.
• Native UAT records known downloaded Cloud matches, offline local search, exact focus/return, account change, guest denial and cache eviction with public Cloud identity/readback evidence.
• No new full-history search service or central message replication. If a public service contract is truly required, publish a separate owner-repo issue/PR with its runnable API/CLI acceptance before depending on it.

Blocked by

• N3: Mobile Search retained Computer messages and restore result context #2196

Delivery and testing boundary

Planning only; do not implement until the Human authorizes this slice. Milestone v2.0; Open Cloud Groups Project. One issue = one implementation PR. UAT only; no Prod/merge authorization. Use existing source owners/public API and CLI readback for business facts, focused behavioral regressions and exact-head native evidence. Reserve the shared device lane; report iOS/Android/physical/Human and large-text/keyboard/accessibility results separately. Preserve unrelated work. Do not infer native success from source tests.