nuget-trusted-publishing
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from repository files to automate workflow generation.
- Ingestion points: Reads metadata from .csproj, Directory.Build.props, .mcp/server.json, and .template.config/template.json in the user repository.
- Boundary markers: None identified; instructions do not specify delimiters or warnings to ignore embedded instructions.
- Capability inventory: Uses dotnet pack for verification and has permissions to create or modify GitHub Actions workflows.
- Sanitization: Does not describe sanitization of extracted strings before use in shell commands or generated files.
Audit Metadata