The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill performs network requests to api.crossref.org, api.openalex.org, and export.arxiv.org to verify citation metadata against official records. These are well-known services in the research and academic community.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and parses bibliographic data (.bib files) provided by the user. While this data is external and untrusted, the script uses a deterministic comparison logic to generate verification reports, minimizing the risk of instructions embedded in the data influencing the agent's behavior.
Ingestion points: The scripts/verify_citations.py script reads and parses BibTeX files via the parse_bib function.
Boundary markers: No explicit instruction delimiters are used when processing the text fields, but the logic is focused on string similarity matching.
Capability inventory: The script performs network GET requests and file writing (via the --json flag).
Sanitization: The norm() function provides basic sanitization by removing LaTeX commands and non-alphanumeric characters for metadata comparison.
[COMMAND_EXECUTION]: The skill requires the agent to execute the provided scripts/verify_citations.py script. The script is self-contained and relies exclusively on the Python standard library.

citation-verification