fix: replace dirs::home_dir() with effective_home_dir() in crates/tui

[EvanProgramming]

Summary

Audit and replace all remaining raw dirs::home_dir() call sites in crates/tui/src with crate::config::effective_home_dir().

Context: Fixed two Windows-only CI test failures in bc461c60c caused by raw dirs::home_dir() calls in tools/workflow.rs and tui/views/skills_manager.rs. Root cause: dirs::home_dir() on Windows resolves via a native OS API call and does not respect HOME/USERPROFILE env var overrides, so tests that fake $HOME via the established EnvVarGuard pattern silently disagreed with production code that used the real Windows user profile instead. crate::config::effective_home_dir() already exists for exactly this (checks HOME, then USERPROFILE, then HOMEDRIVE+HOMEPATH) and is used correctly elsewhere in the crate.

This PR sweeps all remaining call sites (45 replacements across 29 files) so the crate consistently uses effective_home_dir(). The fallback inside effective_home_dir() itself (config/paths.rs:55) is intentionally left as dirs::home_dir().

Note: network_policy.rs and skills/install.rs are #[path]-included from integration tests (tests/skill_cli.rs), so crate::config:: does not exist in the test binary crate context. These two files keep dirs::home_dir() with explanatory comments — they are not test-fakability-critical.

Testing

• cargo fmt --all -- --check
• cargo clippy --workspace --all-targets --all-features --locked — warning-free
• cargo test --workspace --all-features --locked — 8155 passed, 0 failed (only 1 unrelated PTY flake)

Checklist

• Updated docs or comments as needed
• Added or updated tests where relevant (N/A — pure refactor)
• Verified TUI behavior manually if UI changes (N/A)
• Harvested/co-authored credit uses a GitHub numeric noreply address (N/A)

Closes #4757

[github-actions]

Thanks @EvanProgramming for taking the time to contribute.

This repository is observing a maintainer-managed PR intake gate in dry-run mode, so this pull request is staying open. This note helps maintainers prepare the allowlist before any enforcement is considered.

Please read CONTRIBUTING.md for the expected contribution shape. A maintainer can grant recurring PR access by commenting /lgtm on a pull request.

[Hmbown]

Thank you for this, @EvanProgramming — the report and the sweep were right, and #4757 is closed.

Closing as already landed: 0d7312736 landed the same change on main. rg -n 'dirs::home_dir' crates/tui/src/ now returns only 3 hits, all inside crates/tui/src/config/home.rs itself (two doc comments and the OS fallback in the definition), against 78 effective_home_dir call sites.

Credit note: that commit went in without a Co-authored-by trailer for you, which was our miss — your PR predated it by ~11 hours. I'm adding you to .github/AUTHOR_MAP so future harvests in this area attribute correctly.

One thing worth calling out because you got it right and flagged it honestly: you deliberately left network_policy.rs alone with a note that crate::config wasn't reachable from the #[path]-included skill_cli test binary. That was a real constraint, not an oversight. It was later solved by splitting effective_home_dir into a dependency-free leaf (189c93965), so network_policy.rs:324 now uses it too. Your note is what made that the obvious next step.