The Agent Skills Directory

[SAFE]: The skill serves as a knowledge base and instruction set for an AI agent to assist with MSBuild-related tasks. It promotes best practices for .NET build systems and does not contain malicious code, exfiltration logic, or obfuscation.
[DATA_EXPOSURE]: The skill includes explicit warnings in references/binlog-generation.md and references/binlog-failure-analysis.md about the sensitive nature of MSBuild binary logs (.binlog), which can contain environment variables, credentials, and paths. It correctly advises keeping these artifacts local and redacting secrets before sharing.
[COMMAND_EXECUTION]: The skill instructs the agent to use standard .NET development commands such as dotnet build, dotnet msbuild, and PowerShell utilities like Select-String. These are used for legitimate diagnostic purposes (e.g., replaying logs, searching for error codes).
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external build artifacts (project files and binary logs). It mitigates risks by instructing the agent to use MSBuild's official replay tools rather than ad-hoc parsing, and it provides a clear scope in SKILL.md to prevent the agent from performing unrelated tasks like C# refactoring or system-wide migrations.
[PRIVILEGE_ESCALATION]: In references/antipatterns/additional-antipatterns.md (AP-20), the skill specifically warns against platform-specific shell commands and recommends portable built-in tasks, which reduces the risk of accidental privilege abuse or platform-specific failures.

msbuild