The Agent Skills Directory

[SAFE]: The skill's primary function is to provide structured Markdown templates and instructional prompts for academic writing. It does not include any executable scripts, shell commands, or obfuscated logic.
[DATA_EXPOSURE]: The skill includes instructions to read the package.json file in the user's workspace. This is used solely to extract the version number of the agent for inclusion in 'AI Disclosure' sections of manuscripts, which is a standard academic integrity practice and does not involve sensitive credentials.
[EXTERNAL_DOWNLOADS]: The skill integrates with the crossref tool, a well-known and trusted service for retrieving academic metadata and DOI information. This is a legitimate requirement for citation management in scholarly writing.
[INDIRECT_PROMPT_INJECTION]: The skill instructions describe reading local project files like docs/hypothesis.md and results/analysis_summary.json to automate manuscript drafting. While this represents a data ingestion point, the skill does not possess high-privilege tools (such as shell access or dynamic code execution) that would allow for exploitation via malicious data content.

scientific-academic-writing