The Agent Skills Directory

[SAFE]: The skill provides automated utilities for managing academic citations. All identified behaviors, including network requests to academic databases and local file manipulation, are consistent with the primary purpose of the skill.
[EXTERNAL_DOWNLOADS]: The skill connects to well-known academic services to fetch metadata. These include the CrossRef API (api.crossref.org), the NCBI E-utilities API (eutils.ncbi.nlm.nih.gov), and the arXiv API (export.arxiv.org). These are recognized as well-known technology services for research metadata and do not contribute to a verdict escalation.
[INDIRECT_PROMPT_INJECTION]: The skill processes data fetched from external academic APIs, which represents a potential surface for indirect prompt injection if those sources contained malicious instructions.
Ingestion points: Responses from CrossRef, PubMed, and arXiv APIs ingested via scripts/extract_metadata.py, scripts/search_google_scholar.py, and scripts/search_pubmed.py.
Boundary markers: Absent; the skill does not explicitly separate metadata from instructions or warn the agent about untrusted content in paper titles or abstracts.
Capability inventory: The skill possesses file write capabilities (via allowed-tools) to save BibTeX entries and network access to academic endpoints.
Sanitization: The scripts sanitize metadata for BibTeX syntax (e.g., handling page ranges and author separators) but do not filter for natural language instructions. Given the technical nature of the output, the risk is minimal.

alterlab-citation-mgmt