fix staged tailscale installs

[peters]

What changed

• persist bundled offline payloads into .surge-cache during surge setup --stage
• prefer remote app-copy when offline tailscale staging should reduce the next transfer, and check for a matching pre-staged payload before downloading anything locally
• avoid published installers for --stage so remote stage mode always runs a stage-capable surge binary
• replace the new positional stage bool with typed install/stage options and update the affected tests

Why

The previous --stage patch added the flag but left two runtime gaps:

• offline installers only resolved the bundled payload from the extracted installer tempdir, so setup --stage did not leave anything reusable behind once the installer process exited
• remote stage mode could hand --stage to an older published installer whose embedded surge binary did not know that flag, and offline/tailscale staging could still miss the low-transfer activation path on the next install

Impact

• surge setup --stage now produces a durable staged cache entry for offline installers
• surge install --stage --node ... now stages remote tailscale installs in a way the next install can reuse with minimal transfer, especially for offline payloads and persistent-asset updates
• stage mode no longer depends on legacy published installers accepting the new flag

Validation

• ./scripts/sync-surge-core-vendor.sh --check
• ./scripts/check-version-sync.sh
• cargo fmt --all -- --check
• RUSTFLAGS="-D warnings" cargo test --workspace
• cargo clippy --all-targets --all-features -- -D warnings
• cargo clippy --workspace --lib --bins --examples -- -D warnings -D clippy::unwrap_used -D clippy::expect_used
• cargo clippy --workspace --all-targets --all-features -- -D warnings -W clippy::pedantic
• dotnet format dotnet/Surge.slnx --verify-no-changes
• dotnet test dotnet/Surge.slnx --configuration Release