The Agent Skills Directory

[SAFE]: The skill provides a comprehensive and structured workflow for academic research without any malicious patterns detected. All described behaviors are aligned with its stated purpose of facilitating literature reviews.
[COMMAND_EXECUTION]: The documentation includes shell command examples for workflow automation, such as validating file licenses using jq and retrieving paper data using uvx. These commands are transparent, localized to the research directory, and represent standard use of CLI tools.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of academic papers and metadata via the opencite tool. This is a core feature of the literature review process and targets well-known academic resources.
[PROMPT_INJECTION]: As the skill ingests and processes external data (research papers), there is a theoretical surface for indirect prompt injection. However, the workflow effectively mitigates this risk by requiring multi-phase synthesis, human-in-the-loop rigor checklists, and strict citation-to-source traceability.

lit-review