The Agent Skills Directory

[SAFE]: The skill provides a technical framework for engineering quality and pattern recognition without any malicious commands or bypass attempts.
[COMMAND_EXECUTION]: The methodology suggests using grep to audit codebases for patterns such as hardcoded colors or naming violations, which is consistent with the skill's purpose and the allowed-tools configuration.
[SAFE]: The skill explicitly instructs the agent to respect permission boundaries by defining Permission Tripwires that require the agent to stop and explain constraints when encountering protected resources like .env or credential files.

pattern-recognition