The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions involve processing output from the dotnet build command to identify and fix warnings. This creates a potential surface for indirect prompt injection if the project being built contains malicious code designed to produce warnings that influence agent behavior.
Ingestion points: The agent is instructed to read dotnet build output and grep for IL-prefixed warnings in SKILL.md.
Boundary markers: No specific delimiters or safety instructions are provided to sanitize or isolate build output before processing.
Capability inventory: The skill utilizes shell execution (dotnet build, grep) and file editing capabilities (via the agent or suggested sub-agents) as described in SKILL.md.
Sanitization: No input sanitization or validation of the build output content is performed.
[COMMAND_EXECUTION]: The skill utilizes standard development tools for its primary purpose. The commands used (dotnet build, grep) are restricted to identifying compiler warnings and are safe within a development environment.
[REMOTE_CODE_EXECUTION]: No remote code execution or untrusted downloads were detected. The skill provides static C# code snippets for polyfills and serialization contexts, which are standard for the described task.
[CREDENTIALS_UNSAFE]: No hardcoded credentials, secrets, or sensitive file path accesses were identified in the instructions or reference files.

dotnet-aot-compat