citesure 0.5.17 adds ORCID ID validation to catch AI-generated references

citesure 0.5.17: Validating ORCID iDs Before Shipping

Bibliographies are increasingly populated with ORCID iDs. However, the rise of large language models (LLMs) drafting content has introduced a new problem: these models can invent ORCID iDs that appear plausible but are entirely fabricated. The latest release of the citesure tool, version 0.5.17, addresses this by treating ORCID iDs as soft identifiers, similar to how it handles other bibliographic metadata. This means it performs validation checks on their shape, attempts to resolve them against the public ORCID registry, and crucially, warns when a listed ORCID iD clearly does not match the associated author's name.

The primary goal is to prevent the accidental inclusion of incorrect or non-existent ORCID iDs in academic and professional publications. This is particularly important as ORCID iDs become a standard for author attribution and discoverability in research. Fabricated iDs not only undermine the integrity of the bibliography but can also lead to misattribution and confusion for readers and automated systems alike. citesure 0.5.17 provides developers and researchers with a robust tool to preemptively catch these errors.

Key Features of citesure 0.5.17

This release introduces several enhancements to the ORCID validation capabilities:

• Offline Checksum Validation: The command citesure lint-orcids performs an offline validation of ORCID iDs based on the ISO 7064 mod 11-2 checksum algorithm. This allows for rapid checking of ORCID formats without requiring an internet connection, catching basic structural errors immediately.
• Online Resolution and Soft Matching: The command citesure orcid offers a more comprehensive online check. It utilizes the ORCID public API to resolve the provided ORCID iD and retrieve the associated public person record. It then performs a soft-match against the author's name provided in the bibliography. This is the critical step that helps identify when an ORCID iD is valid in format but belongs to a different person than the one listed.
• Disk-Cached Person Records: To improve efficiency and reduce redundant API calls, citesure now implements disk caching for retrieved person records. This means that once an ORCID iD has been resolved and its associated data cached, subsequent checks for the same iD will be significantly faster.
• MCP Tool for Agents: A dedicated command-line tool, citesure_orcid, is available for agents. This utility is designed to streamline the process of checking and validating ORCID iDs within larger workflows or automated systems.
• Domain Pack: The new domain pack, orcid-authors, likely provides a structured way to integrate these ORCID validation functionalities into other applications or workflows, offering a programmatic interface for developers.

How it Works: A Deeper Dive

The process citesure employs for ORCID validation can be broken down into a few key stages, differentiating between offline and online checks. The offline check, citesure lint-orcids, focuses purely on the syntax and mathematical validity of the ORCID identifier. ORCID iDs follow a specific 16-digit format (XXX0-XXXX-XXXX-XXXX) and include a check digit calculated using the ISO 7064 mod 11-2 algorithm. This algorithm ensures that the identifier is numerically sound, much like how a credit card number has a Luhn algorithm check. This initial step is fast and resource-light, catching typos or malformed strings immediately.

The online check, initiated by citesure orcid, goes much further. When you provide an ORCID iD, citesure first performs the offline checksum validation. If that passes, it then queries the ORCID Public API. This API allows programmatic access to publicly available ORCID record information. The tool retrieves the name associated with the ORCID iD from the registry. The crucial step is the subsequent comparison between the name found in the ORCID registry and the author's name as it appears in the bibliography or reference list being processed. The "soft-match" aspect is important here; it implies that the comparison is not a strict, case-sensitive, exact string match. Instead, it likely uses fuzzy matching algorithms or considers variations in spelling, middle initials, and common name orderings to determine if there's a high probability that the ORCID iD and the listed author refer to the same individual. If the match is weak or indicates a discrepancy, citesure flags it as a potential error.