The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external scholarly APIs (arXiv, OpenAlex, Crossref, Semantic Scholar) to generate digests and summaries, creating a potential surface for indirect prompt injection via malicious paper abstracts.
Ingestion points: External data enters the context through multi-source search operations described in SKILL.md and references/gap-analysis.md.
Boundary markers: The documentation does not specify explicit boundary markers or instructions to the agent to ignore embedded commands in the fetched abstracts.
Capability inventory: The skill includes file writing to a local archive (SKILL.md) and network operations via messaging tools for delivery (references/push-format.md).
Sanitization: There is no evidence of explicit sanitization or filtering of external paper content before summarization.
[EXTERNAL_DOWNLOADS]: The skill fetches metadata and abstracts from well-known and trusted academic registries including arXiv, OpenAlex, and Crossref.
[COMMAND_EXECUTION]: The skill utilizes cron for its primary purpose of scheduling daily literature monitoring tasks, as detailed in SKILL.md and references/cron-setup.md. The implementation follows best practices by using project-relative placeholders for paths and identifiers in its documentation.

nature-literature-pipeline