The Agent Skills Directory

[SAFE]: The skill provides maintenance tools and guidelines for documentation hygiene. The analysis found no evidence of malicious behavior, credential exfiltration, or safety bypass attempts.
[COMMAND_EXECUTION]: The skill uses provided Node.js scripts (mb-lint.mjs and mb-doctor.mjs) to validate documentation structure and task records. These scripts operate locally within the repository and are constructed from skill assets rather than external or user-provided input, ensuring a controlled execution environment.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes Markdown and JSON files from the repository to ensure structural compliance. While this processes external data, the analysis is limited to mechanical and workflow validation performed by local scripts, presenting no significant risk of prompt injection.

mb-garden