The Agent Skills Directory

Command Execution: The skill provides instructions to run common developer commands such as dotnet msbuild and grep. These are used to generate and filter diagnostic logs, which is standard practice for build performance analysis.
Indirect Prompt Injection Surface: The skill is designed to analyze external project files and build logs. This ingestion of untrusted data represents a potential surface for indirect prompt injection, although the risk is mitigated by the skill's focus on structural and performance data rather than executing content from the files.
Ingestion points: MSBuild project files (e.g., .csproj) and binary logs (.binlog) as described in the diagnostic procedures in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are defined for the data ingestion process.
Capability inventory: Shell command execution (dotnet msbuild) and data extraction from build artifacts.
Sanitization: No specific sanitization or validation logic is detailed for the content of the ingested build files.

eval-performance