writing-mstest-tests
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a 'Repository overlay' mechanism that loads instructions from a specific file path within the project repository. This creates a surface where malicious instructions embedded in a repository file could influence the agent's behavior.
- Ingestion points:
.agents/skill-overlays/dotnet-test/writing-mstest-tests.md(as described in the 'Repository overlay' section ofSKILL.md). - Boundary markers: The skill requires specific YAML frontmatter fields (
core,binding-revision,mode) to be present and correctly valued as a prerequisite for processing the overlay. - Capability inventory: The agent is authorized to read local project files (
packages.config,global.json,.csproj) and execute thedotnet testcommand-line tool. - Sanitization: No sanitization or filtering of the instructions within the overlay file is mentioned beyond the frontmatter check.
- [COMMAND_EXECUTION]: The skill explicitly directs the agent to execute shell commands to verify that code changes result in passing tests.
- Evidence: Step 9 in
SKILL.mdinstructs: 'run the narrowest affected dotnet test command after editing'.
Audit Metadata