The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown documentation and instructional prompts for academic writing styles (Nature-style) without any executable scripts or command-line operations. \n- [SAFE]: Ethical boundaries (AI traffic-light boundary) are explicitly defined to prevent AI-generated fabrication and ensure human oversight of the scientific argument. \n- [SAFE]: No external network requests or data exfiltration vectors are present in the provided instructions or reference files. \n- [SAFE]: The indirect prompt injection attack surface is evaluated as safe due to a total lack of dangerous tool capabilities (Ingestion points: SKILL.md manuscript input; Boundary markers: structured output format with revision notes; Capability inventory: none; Sanitization: explicit ethics and responsibility guidelines).

nature-polishing