The Agent Skills Directory

[SAFE]: The skill operates entirely within the domain of text analysis and manuscript drafting. It provides instructions for auditing study designs, replication hierarchies, and statistical reporting completeness without requesting access to sensitive files or executing shell commands.
[EXTERNAL_DOWNLOADS]: The skill provides references to official Nature Portfolio domains (nature.com) for reporting standards and submission guidelines. These are well-known and trusted academic resources, and their inclusion is standard for a skill targeting scientific publishing.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided manuscripts, statistical outputs, and reviewer comments. While this presents a vulnerability surface for indirect prompt injection, the skill lacks access to high-privilege tools (such as network-write or code execution) that could be exploited. Furthermore, the use of strict output skeletons and the AUTHOR_INPUT_NEEDED placeholder provides a safeguard against the model being coerced into generating fabricated or unsafe claims. (Severity: LOW).

nature-statistics