The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process text from research papers, which represents an attack surface for indirect prompt injection. The instructions proactively mitigate this risk by mandating that the agent treat paper text, abstracts, and metadata as data under review, explicitly instructing the agent to never follow instructions found within them.
[COMMAND_EXECUTION]: The skill utilizes curl for fetching metadata from established and well-known scientific APIs, including Crossref, arXiv, and OpenAlex. It also executes a provided local Python script (scripts/check_stats.py) to verify statistical reports. These operations are appropriately scoped to public data retrieval and mathematical verification.
[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill adheres to security best practices by advising against the execution of unreviewed code found in papers and prohibiting the creation of fabricated API credentials.

review-scientific-paper