bibcheck verify

    1

    Verify scientific bibliographies against Crossref and OpenAlex to detect hallucinated or incorrect citations.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    bibcheck verify

    Example session with this skill installed

    Check the references in my manuscript draft references.txt and flag any that look like AI hallucinations.

    • Read your context and instructions
    • Compiled the bibcheck verify

    I have analyzed the bibliography in references.txt using metadata extracted from the file and verified it against Crossref and OpenAlex databases.

    Bibliography Verification Report

    The verification process is complete. I processed 12 citations and identified 2 entries that require manual intervention.

    Reference IDStatusIssue
    0verifiedMatches found in Crossref and OpenAlex
    1verifiedMatches found in Crossref
    2suspected_hallucinationNo matching records found in academic databases
    3verifiedMatches found in OpenAlex
    4low_confidenceTitle matches but author list differs significantly
    5verifiedMatches found in Crossref and OpenAlex
    6verifiedMatches found in Crossref
    7verifiedMatches found in OpenAlex
    8verifiedMatches found in Crossref and OpenAlex
    9verifiedMatches found in Crossref
    10verifiedMatches found in OpenAlex
    11verifiedMatches found in Crossref and OpenAlex

    The results have been saved to the output directory. Reference

    ID 2 shows no indexed record for the provided title and author combination, suggesting it may be fabricated or contains significant typos. Reference

    ID 4 returned a partial match but the metadata confidence is below the threshold for automatic verification.

    Next steps

    1. Manually verify the existence of the paper at ID 2 using a search engine.
    2. Review the metadata for ID 4 to ensure the DOI and author names are correctly formatted.
    3. Update the references.txt file with the correct details for flagged entries and re-run the check.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Scientific papers and technical reports often contain fabricated or hallucinated citations that are difficult to spot manually. This skill automates the verification of bibliographies by cross-referencing metadata against Crossref and OpenAlex databases. It converts raw citations from PDFs, Markdown, or BibTeX into structured queries and flags high-risk entries for manual review.

    What it does

    • Multi-format extraction identifies citation metadata from PDF, Markdown, text, or BibTeX files.
    • Automated verification cross-references extracted titles, authors, and DOIs with Crossref and OpenAlex.
    • Hallucination detection flags references that cannot be found in major academic indexes.
    • Triage reporting generates a summary of verified, low-confidence, and suspected-hallucination citations.
    • Metadata normalization structures fragmented citation data into a standardized JSON format for processing.

    How it works

    1. The agent reads the source bibliography and extracts metadata for every citation, preserving the original order.
    2. It generates a temporary JSON file containing structured titles, authors, years, and search queries.
    3. The bibcheck-verify utility runs against the original file using the structured metadata as a reference.
    4. The skill parses the resulting triage report to identify which citations require human verification.

    Frameworks & tools

    This skill requires Python 3.11+ or the uv package manager. It utilizes the bibcheck-verify Python package and connects to the Crossref and OpenAlex public APIs via HTTPS.

    Why this beats prompting it yourself

    LLMs are prone to confirming their own hallucinations when asked to verify text. This skill forces the agent to check external, authoritative academic databases, providing evidence-based verification rather than probabilistic guesses.

    Use cases

    • Peer review to verify that a submitted manuscript contains legitimate and accurately cited references.
    • Academic writing to ensure your own bibliography lacks accidental hallucinations or broken DOIs.
    • Technical due diligence to validate the sources cited in whitepapers or technical reports.

    Known limitations

    Requires outbound HTTPS access to api.crossref.org and api.openalex.org. Triage labels like "suspected_hallucination" are not definitive proof of fabrication and always require human review.

    Requires Python 3.11+ and outbound HTTPS to Crossref and OpenAlex APIs. Only supports metadata fields for title, authors, year, doi, and venue.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Python
    uv (Python tool runner)

    Permissions

    Terminal / Shell
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    The skill identifies citations in your document and cross-references them with Crossref and OpenAlex databases. It flags references that have no record in these databases, helping you catch LLM-generated hallucinations or human errors that are difficult to spot manually.

    Yes, the skill supports PDF, Markdown, BibTeX, and plain text files. It extracts citation metadata from these formats and converts them into a structured JSON format for automated verification against academic indexes.

    This skill requires a Python 3.11+ environment and outbound HTTPS access to connect to the Crossref and OpenAlex APIs. It is designed to work with any AI agent capable of executing Python-based tools and managing file I/O.

    You will receive the complete skill package including the integration logic, documentation, and the necessary configurations to connect to the bibcheck-verify utility and external academic APIs.

    While a high-risk flag suggests a hallucination, it can also occur for very new publications not yet indexed or private internal documents. The skill provides a triage report that narrows down which citations require your manual oversight, saving time on the ones that are verified.

    New in Research & Analysis