The Agent Skills Directory

[COMMAND_EXECUTION] (MEDIUM): The skill's primary function is to execute local scripts and commands. Specifically, the documentation states it will automatically run dotnet run from the ./Umbraco-CMS.Skills/ directory. This automated process-spawning capability could be exploited if the targeted project directory contains malicious code.
[CREDENTIALS_UNSAFE] (LOW): The skill documentation explicitly requests sensitive environment variables UMBRACO_USER_LOGIN and UMBRACO_USER_PASSWORD. While these are not hardcoded, their usage within a script (run-tests.ts) that is not provided for inspection creates a risk of logging or accidental exposure of credentials.
[INDIRECT_PROMPT_INJECTION] (LOW):
Ingestion points: The skill processes "skill examples" which are external files provided to the agent.
Boundary markers: Absent. The documentation does not specify delimiters or instructions to ignore embedded prompts within the test examples.
Capability inventory: The skill uses Bash, Read, and Glob tools and can execute arbitrary shell commands.
Sanitization: Absent. There is no evidence that the test runner sanitizes the output of the tests before generating the JSON report, which the agent might later consume.

umbraco-skill-test-runner