Audit research implementations against published papers across five fidelity dimensions.
npx skills add https://github.com/Borda/AI-Rig --skill verify-borda
Please help me install this Agent Skill. Skill: verify Source: https://github.com/Borda/AI-Rig/tree/main/plugins/research/skills/verify Command: npx skills add https://github.com/Borda/AI-Rig --skill verify-borda
Audits research implementations to ensure code faithfully implements published paper methods across five evaluation axes.
Invoke the verify workflow with a paper input, for example: /research:verify <paper.pdf|arxiv-url|'pasted text'>.
High-intent search queries and answers about installing and using this skill.
To verify code against a research paper, you audit the implementation across five dimensions: formula matching, hyperparameter parity, evaluation protocol, notation consistency, and citation chain. This structured verification extracts claims from the paper and aligns them with code references to produce a fidelity report.
Reproducibility verification in machine learning research is the process of auditing an implementation to ensure it faithfully executes the methods described in a published paper. It involves checking formula matching, hyperparameter parity, and evaluation protocols to confirm experimental alignment.
You audit hyperparameters and formulas by extracting specific claims from the paper text and aligning them with repository code references. The audit generates a structured claims table and remediation notes to guide fixes for any hyperparameter parity or formula matching discrepancies.
Yes, you can use an arxiv URL to check code fidelity. The verification workflow accepts paper inputs as PDF files, arxiv URLs, or pasted text to extract claims and compare them against your repository code across five evaluation axes.
A limitation of experiment verification is that it primarily targets research workflows where you have already implemented a method from a paper. It requires both the paper content and repository code to produce a structured verification report, meaning incomplete implementations or inaccessible paper text may limit the audit.
Yes, the verify workflow generates a structured claims table and remediation notes for code alignment issues. It identifies discrepancies across formula matching, hyperparameter parity, evaluation protocol, notation consistency, and citation chain to guide your fixes.
ShiienExtract paper metadata and core insights into structured markdown notes.
drpedapatiRecord experiment provenance and reproducible run metadata for research workflows.
mvillmowExtract LaTeX equations from research papers and map components to code.
mkurmanEvaluate manuscripts and grant proposals using standardized peer review checklists.
