triage
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- Dynamic Code Generation and Execution: The skill is designed to generate a C# console application from external issue descriptions and execute it to confirm bugs. This involves running code derived from untrusted user input, which is a significant security consideration in any execution environment.
- Indirect Prompt Injection Surface: The skill ingests data from external GitHub issues which is then used to influence code generation and triage decisions. This creates a surface where malicious instructions embedded in an issue could potentially affect the agent's behavior.
- Ingestion points: Input issue text and linked resources (provided in SKILL.md).
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing the issue content.
- Capability inventory: The skill utilizes the GitHub CLI (gh), file system access to create and run .NET applications, and potentially network or container access for database providers (SKILL.md).
- Sanitization: There is no mention of explicit sanitization or validation of the user-provided code snippets or text before they are integrated into the reproduction template.
- Access to Sensitive Environment Variables: The reproduction template accesses the Test__SqlServer__DefaultConnection environment variable. While standard for automated testing in the dotnet/efcore repository, this variable contains sensitive connection string information.
- Sensitive Data Logging: The template includes the EnableSensitiveDataLogging() configuration. This is a standard EF Core feature for debugging but means that the generated logs—which the skill is instructed to post back to GitHub—could contain literal data values from the test execution.
Audit Metadata