fix(daemon): restore pooled connection timeout after short-timeout operations

[Wirasm]

Summary

• Four functions (ping_daemon, get_session_status, get_session_info, read_scrollback) set a 2s read timeout on pooled IPC connections but never restored the 30s default before returning to the pool — the next caller on the same thread inherited the corrupted timeout
• Save original timeout before override, restore before pool return; if restore fails, drop the connection instead of poisoning the pool
• Add IpcConnection::get_read_timeout() getter to support the save/restore pattern

Test plan

• New unit test test_get_read_timeout_returns_current_value verifies round-trip save/set/restore
• cargo fmt --check passes
• cargo clippy --all -- -D warnings passes
• cargo test --all passes (all 27 test suites)
• cargo build --all passes

[Wirasm]

PR Review Summary

Critical Issues (0 found)

None.

Important Issues (1 found)

Agent	Issue	Location
code-reviewer	The save/restore pattern is repeated 8 times across 4 functions. The same crate already has a TimeoutGuard RAII struct (kild-protocol client.rs:75-84) that solves timeout restoration for is_alive(). A similar RAII approach or a with_short_timeout helper would eliminate the boilerplate and guarantee restoration even on panics or early returns. The current approach is correct for a targeted bug fix, but consider a follow-up to DRY this up.	crates/kild-core/src/daemon/client.rs (lines 344-348, 394-398, 421-423, 461-465, 475-477, 571-575, 586-588, 599-601)

Suggestions (2 found)

Agent	Suggestion	Location
code-simplifier	The TimeoutGuard RAII pattern from is_unix_alive() could be lifted into a public helper on IpcConnection to make the save/override/restore cycle a one-liner for callers. This would also make the restore unconditional across all match arms (including the Ok(unexpected) arms that currently just drop the connection).	crates/kild-protocol/src/client.rs:75-84
type-design-analyzer	get_read_timeout() is well-designed -- mirrors set_read_timeout() symmetrically, handles both stream variants, and matches the stdlib return type exactly. No changes needed.	crates/kild-protocol/src/client.rs:244-251

Strengths

• Correctly identifies a real pool-poisoning bug: short timeouts leaking into pooled connections
• The "drop instead of return" fallback when restore fails is exactly right -- fail-safe, never poisons the pool
• New test test_get_read_timeout_returns_current_value verifies the full save/set/restore round-trip
• Consistent application of the fix across all 4 affected functions
• PR description clearly explains the root cause and fix strategy

Verdict

READY TO MERGE -- The fix is correct, consistent, and well-tested. The DRY concern is real but appropriate as a follow-up, not a merge blocker for a targeted bug fix.

Recommended Actions

1. Consider a follow-up PR to extract the save/restore pattern into an RAII helper (reusing the existing TimeoutGuard concept), reducing the 8 repetitions to 4 one-liners
2. No blocking changes needed