fix: harden sync and failure handling

[AksharP5]

Repository and GitHub failures could leave Blippy with stale or misleading state. Scans could erase cached remotes, linked probes could retry continuously, cross-repo links could jump to the wrong item, and PR metadata failures appeared as valid empty data.

This keeps cached data intact, makes unavailable metadata explicit, and stops failed background probes from hammering the API. It also includes the earlier sync, cache, discovery, CLI, input, and CI hardening from this branch.

What changed

• preserve cached remotes when Git inspection fails
• stop automatic linked-item retries after API failures
• reject cross-repository links that cannot be opened safely in the current TUI context
• show unknown PR view state and surface review metadata failures
• close issues before posting optional close comments to avoid duplicate comments on retry
• keep sync snapshots, repository discovery, input handling, CLI behavior, and CI failure paths consistent

Checks

• cargo test --all-targets --all-features --locked
• cargo clippy --all-targets --all-features --locked -- -D warnings
• cargo fmt --all -- --check
• git diff --check