The Agent Skills Directory

Command Execution through MSBuild: The skill constructs shell commands to invoke the repository's dotnet tool. This is necessary to evaluate complex test environment metadata. The skill mitigates potential injection by validating requested scenario names against the repository's project files before execution.
Indirect Prompt Injection Surface: The skill ingests data from local project files, which could theoretically contain malicious instructions intended to influence the agent's behavior.
Ingestion points: src/tests/Common/testenvironment.proj and specified .csproj files.
Boundary markers: The skill lacks explicit prompt boundary markers for interpolation but uses specific parsing rules for shell assignments.
Capability inventory: Execution of shell commands via dotnet msbuild and file system operations (write and delete).
Sanitization: Validation of scenario identities and targeted parsing of script output (set or export assignments) minimizes the processing of unintended or malformed data.
Temporary File Handling: The skill generates temporary scripts to capture MSBuild output. It includes logic to parse these files for specific assignment patterns and ensures the temporary files are deleted after the conversion process is complete, reducing the risk of persistence.
Information Disclosure Prevention: The skill includes explicit instructions to avoid printing sensitive values or CI credentials in the final response. This design helps prevent the accidental exposure of environment secrets during the reporting phase.

test-scenario-env