METHODOLOGY / WORKING PRINCIPLES
Show the basis. Keep the uncertainty.
We are designing an assessment standard for developer and automation work. This page describes the intended method, not a published rubric, approved level or active certification process.
01 / SCOPE
Start with the project problem.
Record the business use case, connected systems, delivery stage and outcome evidence. A CRM lead-routing prototype and a production workflow should remain distinguishable in both assessment and employer matching.
02 / CONTRIBUTION
Ask what the person actually did.
Separate self-reported work, observable artifacts, AI-generated material and collaborators' contributions. Selected chats can illuminate decisions and revisions when available, but missing conversation history must not become a negative skill judgment.
03 / QUALITY
Look for evidence of working well.
Relevant tests, handling of failures, maintainability, documentation and deployment evidence may matter. The exact criteria and weights must be defined and calibrated before a level can be awarded.
04 / REVIEW
Make conclusions challengeable.
Each finding should point to its basis and say what remains unknown. Candidates need a way to correct context or dispute a claim. A reviewed conclusion must not be presented as the same thing as a machine inference.
05 / VERIFICATION
Keep the issued record stable.
A future verifier should show the standard version, assessed scope, date and current credential status. Later profile additions can add context, but cannot silently change what was originally assessed.
NEXT READING