The Agent Skills Directory

[SAFE]: No malicious patterns, command execution, or unauthorized network operations were detected. The skill serves as an instructional framework for text-based critique and improvement using clearly defined evaluation prompts.
[INDIRECT_PROMPT_INJECTION]: The skill operates by processing external content for review, which constitutes a standard surface for indirect prompt injection. Ingestion occurs in the evaluation prompts where user content is assessed. No explicit boundary markers or sanitization are defined, but the skill's capabilities are limited to generating markdown reports, presenting minimal risk.
[METADATA_POISONING]: The skill contains platform-specific metadata such as 'governance_phases' and 'organ_affinity'. These fields are consistent with the vendor's documented environment and do not include deceptive or malicious instructions.

evaluation-to-growth