The Agent Skills Directory

[SAFE]: The skill provides a robust framework for time-constrained task execution, utilizing fresh clock anchors and protected validation reserves to ensure agent accountability. Admission rules prevent execution if bounds are unavailable or estimates are insufficient.
[SAFE]: Local Python scripts (e.g., scripts/validate_skill.py, scripts/check_code_rules.py) enforce structural and stylistic standards using standard library modules like ast and json. These tools are used for development-time validation and do not perform network or disk operations outside of the skill package.
[SAFE]: Task automation via mise.toml is restricted to local scripts and unit tests, following best practices for environment-controlled execution without external dependencies or remote script downloads.
[SAFE]: The skill contains specific defensive instructions to prevent prompt injection from overriding its control logic, such as the directive to not replace instructions with remembered text.

timebox