The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill performs network operations to fetch scholarly metadata from api.crossref.org and DOI registration agency information from doi.org. These are well-known academic services essential for the skill's bibliographic functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes bibliographic records retrieved from the external Crossref API, which represents an attack surface for indirect prompt injection.
Ingestion points: Bibliographic metadata (titles, abstracts, reference lists) is ingested from api.crossref.org via the CrossrefAPI class in scripts/crossref_api.py.
Boundary markers: Boundary markers or explicit instructions to ignore embedded commands are not present in the code that handles retrieved API data.
Capability inventory: The skill's capabilities are limited to standard network requests (urllib.request) and metadata processing; it does not utilize subprocesses, file-write operations, or dynamic code execution on the ingested data.
Sanitization: The skill performs sanitization of abstract metadata by stripping JATS XML tags and unescaping HTML entities in the clean_abstract function within scripts/crossref_api.py.
[SAFE]: The skill does not contain obfuscated code, credentials, or persistence mechanisms. The implementation relies solely on the Python standard library, requiring no external third-party dependencies.

crossref-search