The Agent Skills Directory

[COMMAND_EXECUTION]: The aot-test-runner.fsx script uses System.Diagnostics.Process to programmatically execute dotnet build and dotnet publish commands. It further executes the resulting compiled binaries to perform 'smoke tests' using --version and --help arguments.
[EXTERNAL_DOWNLOADS]: The included F# diagnostic and analysis scripts (aot-analyzer.fsx, aot-diagnostics.fsx, and aot-test-runner.fsx) utilize the #r "nuget: ..." directive to download and load the System.Text.Json and Argu packages from the official NuGet registry at execution time.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted external data, specifically build logs and source code, and provides summarized reports to the agent.
Ingestion points: aot-analyzer.fsx reads build log files; aot-diagnostics.fsx recursively scans project source files (.cs, .fs).
Boundary markers: None identified; input data is processed as raw text without specific delimiters or instructions to ignore embedded commands.
Capability inventory: The skill has the ability to read the file system, execute dotnet CLI commands, and run generated executables.
Sanitization: There is no evidence of input sanitization or validation of the content within logs or source files before analysis.

aot-guru