The Agent Skills Directory

[SAFE]: The skill instructions and associated scoring rubric follow best practices for narrative evaluation. No malicious patterns, obfuscation, or unauthorized data access were detected. All file operations and tool usages are consistent with the documented purpose of narrative quality assurance.\n- [PROMPT_INJECTION]: The skill processes untrusted narrative content provided via the --source-path parameter. While this represents an indirect prompt injection surface, the risk is mitigated by strict execution protocols and constraints that enforce a read-only review process and direct the agent to perform an objective analysis based on specific metrics.

narrative-review