The Agent Skills Directory

[SAFE]: The skill provides software engineering guidelines and best practices in markdown format, such as 'Laziness Protocol', 'Boundary Discipline', and 'Fix Root Causes'. It does not contain any executable code, network operations, or hardcoded credentials.
[SAFE]: The instructions regarding agent autonomy ('Never Block on the Human') include explicit safety guardrails that require human confirmation for irreversible or high-risk actions like data deletion.
[SAFE]: No scripts, external URLs, or third-party dependencies are included in the skill. All references point to local documentation within the skill's own directory structure.

principles