The Agent Skills Directory

[SAFE]: The skill follows a highly structured and professional academic workflow.
[NO_CODE]: The skill consists of markdown-based instructions for AI agents and does not ship with executable binaries or malicious scripts.
[DYNAMIC_EXECUTION]: The visualization_agent generates and potentially executes Python or R code to create charts. While this involves runtime code generation, it is the primary purpose of the agent and uses well-known libraries (matplotlib, ggplot2) for data visualization.
[EXTERNAL_DOWNLOADS]: The skill mentions external tools like pandoc, tectonic, and databases like Retraction Watch, which are well-known and standard in the academic domain.
[INDIRECT_PROMPT_INJECTION]: The skill handles untrusted data (reviewer comments), but it includes a 'Writing Quality Check' and 'Anti-Leakage Protocol' to ensure the AI remains focused on the provided research materials rather than being influenced by external instructions.

academic-paper