The Agent Skills Directory

[SAFE]: The skill serves as a purely educational and instructional resource for Rust performance optimization. It contains no executable malicious code, obfuscation, or exfiltration patterns.
[COMMAND_EXECUTION]: All shell commands provided in the documentation (e.g., cargo build --release, perf record, strace, valgrind, hyperfine) are standard industry-standard tools for software development, profiling, and performance measurement.
[UNSAFE_USAGE]: The skill includes guidelines for using Rust's unsafe keyword (e.g., get_unchecked, MaybeUninit, transmute). These are correctly presented within the context of performance engineering, including explicit warnings about Undefined Behavior (UB) and documentation of safety invariants.
[DATA_EXPOSURE]: No hardcoded credentials, sensitive file paths, or network exfiltration attempts were found. File paths mentioned (e.g., /tmp/mtrace.log, state.json) are used as placeholders for local development tasks.
[OBFUSCATION]: A thorough review of the extensive prose and code examples found no evidence of Base64 encoding, hidden zero-width characters, homoglyph attacks, or steganographic URL embedding (acrostics).

rust-performance-best-practices