Self-hosted alternatives to datanalysis-credit-risk

D

datanalysis-credit-risk

Credit risk data cleaning and variable screening pipeline for pre-loan modeling

datanalysis-credit-risk is an agent skill from the github/awesome-copilot collection for cleaning raw credit data and screening its variables before a pre-loan model is built. It runs a pipeline of 11 steps, each executed without deleting the original data, so the raw file is always kept. The pipeline assesses data quality, handles missing values, and screens the candidate variables so that only the useful and stable ones go into the model. A quick start runs the complete pipeline from the bundled example script, and the skill documents each step so an analyst can see what was done to the data and why. It applies when working with raw credit data that needs quality assessment, missing-value treatment and variable selection for a credit risk or pre-loan model. Credit models affect whether people get loans, so the choices the pipeline makes, such as how missing values are filled and which variables are dropped, must be justified and reviewed by the model owners, and the model should be checked for fairness and regulatory requirements in the jurisdiction where it will be used. The skill is installed with the skills CLI from the github/awesome-copilot repository. Its skills.sh page showed about 7.3K installs, with Gen Agent Trust Hub and Socket showing warnings and Snyk passing at the time of the scan.