To ensure that the code we publish is appropraite, clear and reusable we require a set way of working when pushing code to private and public repoistories. It is difficult to set any precise but generic standards and so we ask that the following checklist is considered with mandatroy items filled out.
- Does your code have an appropriate licence and copyright notice? (Mandatory)
- Is there a README and does it document intended purpose? (Mandatory)
- Is the README clear and concise? (Optional - use analytics unit template or example)
- Do you need to consider MHRA 'software as a medical device' guidance? (Mandatory - use flowchart)
- Who has responsibility for ongoing support and communications for the code? (Optional Best practice to assign, else state the code is not maintained and when the last update was)
- Has a responsible disclosure process for security issues been defined? (Mandatory - could use github issues)
- Who will address identified issues and security concerns? (Mandatory. - assign monitoring of above disclosure process)
- Has semantic versioning been used? (Optional)
- Have contribution guidelines or PR style guides been included? (Optional)
- Are package dependencies and libaries documented with versions? (Optional)
- Does the code include any sensitive, personal, secret or top secret data/information? (Mandatory)
- Does the code include any unreleased policy? (Mandatory)
- Does the code include business sensitive algorithms (e.g. finance allocations)? (Mandatory)
- Has written permission been obtained for any data stored from the data owner? (Mandatory)
- Are any data transfers conducted safely and securely? (Mandatory)
- Are any credentials contained in the source code? (Mandatory - check in both current version and git history)
- Are any secret keys contained in the source code? (Mandatory - check in both current version and git history)
- Are the commit messages informative? (Optional)
- Do the commit messages include any sensitive information (e.g. names)? (Mandatory)
- Does the git history contain any sensitive information (e.g. at one time real data or credentials were in the code but have since been removed) (Mandatory)
- Have notebook outputs been removed/checked for sensitive information? (Mandatory - check but some appropriate outputs maybe useful: Example)
- Is configuration written as code and separated from analytical code? (Optional)
- Is the code version controlled using GIT or similar? (Optional)
- Is the code stored in your organisational GitHub account? Is it the same organisation that funds the relevant staff time? (Optional)
Any third-party tools you use to host or manage your code follow the National Cyber Security Centre’s cloud security guidance
- Are third party tools used within the code? (Mandatory check. Best practice is to keep an inventory)
- If so do they adhere to the NCSC's Cloud Security Principles? (Mandatory)
- Has a colleague reviewed the code for sensitive data content and security vulnerabilities? (Mandatory - Best practice is to record automated code quality and security tools used)
- Has a code quality review been completed focussing on the end usability and clarity? (Optional - consider runing through the example or similar code quality checklist)
- Has the code been assessed for its level or RAP(Reproducible Analytical Pipeline)?