The Agent Skills Directory

[SAFE]: The skill follows secure patterns for handling academic content. It uses a structured routing system to load instructions from local disk fragments based on the detected input format. No obfuscation, unauthorized network operations, or malicious command execution patterns were found.
[INDIRECT_PROMPT_INJECTION]: The skill processes external text from academic papers, which is a standard surface for indirect prompt injection. However, the skill implements strict structural requirements (original/translation pairs and source-map anchors) and includes instructions to translate for meaning, not style, which serves as a mitigating factor for accidental obedience to embedded instructions.

nature-reader