The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on a custom CLI tool sdd and the Bash tool to perform test discovery, execution, and analysis. This is standard behavior for a development-focused skill.
[DATA_EXPOSURE]: Debugging information, including error messages and code snippets, is shared with external AI services (Gemini, Codex, Cursor) to facilitate root-cause analysis.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) because it ingests untrusted data from test outputs and source code.
Ingestion points: Test logs, stack traces, and source code files are processed by the agent during investigation and consultation phases.
Boundary markers: The instructions do not specify the use of delimiters or markers to distinguish between system instructions and data from external logs or code.
Capability inventory: The agent has access to shell execution (Bash) and file modification tools (Edit, Write).
Sanitization: There is no mention of sanitizing or filtering the content of test logs or source code before analysis.

run-tests