⚡ Bolt: Optimize format_security_list memory allocations - #136

💡 What:
Updated format_security_list in src/config.rs to accept a &mut String buffer instead of returning a new allocated String. Replaced intermediate Vec allocations and format! macro calls in the loop with direct buffered appending via std::fmt::Write::write! and push_str. Updates were made to to_formatted_string and unit tests to pass the buffer as expected.

🎯 Why:
The original implementation created a new Vec of String components (ref = ..., tag = ..., etc.) and used parts.join(", ") inside a loop for each entry, returning a fully allocated String concatenated to the root TOML buffer. This caused O(N) heap allocations when writing large vetted or compromised security lists, introducing unnecessary overhead during configuration serialization.

📊 Impact:
Significantly reduces memory allocations and CPU overhead when saving configurations with extensive vetted or compromised arrays by streaming string data directly into the final toml_str allocation block.

🔬 Measurement:
Run cargo bench (if set up) or profile memory allocations when writing configurations containing large SecurityEntry lists. Run cargo test to ensure serialization integrity.

PR created automatically by Jules for task 5574614858700912070 started by @ffalcinelli