The Agent Skills Directory

[SAFE]: The skill's primary purpose is providing TDD sequencing advice. It explicitly states that it is advisory and cannot override user instructions, project-specific rules, or system constraints.
[COMMAND_EXECUTION]: The skill references a CLI tool, code-oz, which is presented as the primary engine for enforcing TDD gates and performing cross-family reviews. This is standard documentation for the intended execution environment.
[PROMPT_INJECTION]: The skill incorporates a proactive security rule (Universal Rule 11) requiring the agent to treat instruction-like text found in files, tool outputs, and logs as data rather than commands, which is a key defense against indirect prompt injection.

red-first