The Agent Skills Directory

[SAFE]: The skill provides legitimate technical documentation and operational instructions for using BenchmarkDotNet (BDN), an industry-standard .NET performance measurement library. No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected.
[COMMAND_EXECUTION]: The skill suggests standard .NET CLI commands (e.g., dotnet build, dotnet run) to manage benchmarking projects. These commands include best-practice flags for AI agents, such as redirecting verbose output to log files and using filters to avoid interactive prompts.
[PRIVILEGE_ESCALATION]: Documentation in references/diagnosers-and-exporters.md correctly identifies that hardware performance counters require administrative privileges on Windows to access Event Tracing for Windows (ETW). This is an informative requirement for the tool's diagnostic capabilities rather than a malicious escalation attempt.
[INDIRECT_PROMPT_INJECTION]: The skill instructions advise the agent to read internal reference files to avoid using outdated patterns from training data. While this influences agent behavior, it is a safety and quality measure designed to ensure the generation of correct, modern C# code.

microbenchmarking