platform-detection
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute
dotnet --versionto identify the .NET SDK version in the current environment. This is a standard, non-destructive diagnostic command used to ensure the accuracy of the platform detection logic. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from project configuration files, creating a surface for indirect prompt injection.
- Ingestion points: The agent reads
global.json,.csproj,packages.config,Directory.Build.props,Directory.Build.targets, andDirectory.Packages.propsfrom the local repository as specified in SKILL.md. - Boundary markers: The instructions include a specific "Response contract" and "scope gate" to structure the output, but they do not explicitly define security delimiters for the content read from these files.
- Capability inventory: The skill's capabilities are limited to reading local files and checking the SDK version. It does not utilize tools for network access, file writing, or administrative system changes.
- Sanitization: No explicit sanitization of the input files is performed; however, the risk is mitigated because the content is used for property resolution and classification rather than being treated as executable instructions.
Audit Metadata