The Agent Skills Directory

[SAFE]: The skill uses a role-playing persona ('Test Terminator') to enforce a strict and thorough code review process. While the language is aggressive, it is used to define the audit's scope and quality rather than to bypass safety protocols.
[DATA_EXPOSURE]: The skill instructions involve searching and analyzing the user's codebase (e.g., using grep or file path mapping). This is standard and necessary functionality for a code review tool and does not involve external data exfiltration.
[PROMPT_INJECTION]: There are no attempts to override system prompts, extract internal instructions, or bypass safety filters. The instructions focus entirely on the technical review of provided code and requirements.
[COMMAND_EXECUTION]: The installation examples in the README use standard npx commands for adding local or remote skills, which is typical for the supported platform's ecosystem.

test-terminator