The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from untrusted repository files to determine its execution logic.
Ingestion points: According to the 'Bootstrap When Missing' section in SKILL.md, the skill reads project-specific files such as *.csproj, Directory.Build.*, and .editorconfig to detect the target framework and language versions.
Boundary markers: The skill does not define explicit delimiters or instructions for the agent to disregard potential instructions hidden within the metadata of these files.
Capability inventory: As noted in the INVOKES section of SKILL.md, the skill has the capability to inspect repository context, edit files, and execute dotnet CLI commands (build, test, validation).
Sanitization: There is no evidence of sanitization or validation of the project file content before it is incorporated into the agent's reasoning or subsequent file edits.

modern-csharp