The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted academic data and web search results, creating a potential surface for indirect prompt injection attacks.
Ingestion points: Data is ingested from external sources including Semantic Scholar, arXiv, Tavily, Exa, AMiner, and Gemini search results (SKILL.md).
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands in the retrieved abstracts are provided.
Capability inventory: The skill executes local search scripts, accesses the network via search tools, and can trigger other skills like zotero-management.
Sanitization: No sanitization of the retrieved external content is specified.
[COMMAND_EXECUTION]: The skill defines workflows that execute local search scripts with user-supplied query arguments.
Evidence: Use of node scripts/search/semantic-scholar.mjs, node scripts/search/arxiv.mjs, and node scripts/search/search.mjs.

literature-search