The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/los_code_gate.py invokes local git commands using subprocess.run to identify changed files for analysis. This is a benign implementation used solely for repository state inspection.\n- [PROMPT_INJECTION]: The skill includes defensive mechanisms in scripts/codex_hooks/user_prompt_submit.py which use regular expressions to detect and block prompts that attempt to subvert security guidelines or request the hardcoding of secrets.\n- [SAFE]: The skill provides various templates and scripts (e.g., adr_lint.py, arch_law_check.py) that perform local static analysis on software design documents and code. These tools are designed to identify quality and security omissions based on established engineering principles.

laws-of-software-los