fix: propagate linter exit code in rtk lint

[polaminggkub-debug]

Summary

• rtk lint was always returning exit code 0 regardless of the underlying linter's exit status
• Now propagates the linter's actual exit code using !output.status.success() pattern
• Handles both non-zero exit codes (ESLint 1/2, ruff, pylint, mypy) and signal termination (SIGINT/SIGTERM)

Root Cause

The run() function in lint_cmd.rs captured the exit code for tee output but never used it to set the process exit status — it always returned Ok(()) (exit 0).

Fix

Added 3 lines after tracking, matching the exact pattern used by every other RTK module (git, cargo, playwright, vitest, etc.):

if !output.status.success() {
    std::process::exit(output.status.code().unwrap_or(1));
}

Why !status.success() instead of exit_code != 0

Per Rust stdlib docs, ExitStatus::code() returns None when a process is terminated by signal. !success() correctly detects failure in all cases (non-zero exit AND signal termination), while != 0 depends on the unwrap_or fallback value.

Test plan

• All 7 existing lint_cmd tests pass (cargo test lint_cmd)
• Manual test: rtk lint npx eslint <file> --max-warnings=0 returns non-zero exit code
• Verify: rtk lint npx eslint <clean-file> still returns exit code 0

Fixes #185

🤖 Generated with Claude Code

[polaminggkub-debug]

Note on CI validate failure: This is a pre-existing issue on master, not related to this PR.

The scripts/validate-docs.sh check fails because Cargo.toml is at version 0.22.0 but README.md, CLAUDE.md, and ARCHITECTURE.md still reference 0.20.1. This affects all PRs against master currently.

This PR only touches src/lint_cmd.rs (4 lines added).