dotnet-review-inputs
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes .NET source code and project metadata, which creates an indirect prompt injection surface.
- Ingestion points: Generator source code, project files (.csproj, .fsproj), and package manifests (SKILL.md).
- Boundary markers: No explicit delimiters are specified to isolate untrusted data from the agent's instructions.
- Capability inventory: The agent is limited to file system reads and report generation; instructions explicitly forbid modifying code.
- Sanitization: No sanitization or filtering of ingested content is described.
- [EXTERNAL_DOWNLOADS]: The skill references documentation for official .NET testing frameworks.
- Evidence: Links to fscheck.github.io and hedgehogqa.github.io in references/exploration.md.
Audit Metadata