The Agent Skills Directory

[SAFE]: No malicious patterns or behaviors were identified. The skill's instructions establish a clear priority hierarchy where user-provided instructions (e.g., in CLAUDE.md) explicitly override the skill's own directives.
[NO_CODE]: The skill consists entirely of instructional markdown and reference files (LICENSE, tool mappings). There are no executable scripts, binaries, or automated shell commands present in the analyzed content.
[PROMPT_INJECTION]: While the skill uses strong imperative language ("ABSOLUTELY MUST", "not negotiable") to enforce its workflow, these are behavioral directives aimed at improving agent consistency rather than attempts to bypass security filters or extract system prompts.

using-superpowers