The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill imports execution instructions and protocols from an external repository: literature-engine/skills/literature-search/SKILL.md. This introduces a dependency on third-party code from a source that is not recognized as a trusted provider or part of the author's defined infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill is specifically designed to ingest and analyze academic papers, which serves as a surface for indirect prompt injection if those papers contain hidden instructions.
Ingestion points: Untrusted paper content is retrieved and processed through the get_paper_content requirement.
Boundary markers: The skill lacks specified delimiters or instructions to treat the paper content as potentially adversarial data.
Capability inventory: The agent is directed to generate depth-level reports and summaries based on the ingested content.
Sanitization: No sanitization or validation steps are described to filter out malicious patterns within the academic literature.

paper-search