Verification Rubrics

Status: The standard, applied to every artifact.

Purpose

Every module ends with an artifact + verification step. These rubrics define what "good enough" looks like so evaluation stays consistent and honest.

Core Rubric (use on almost every artifact)

Criterion	Weak	Acceptable	Strong
Clarity	Vague, jargon-heavy, or padded	Understandable to a competent outsider	Precise, plain, no wasted words
Grounding	Relies on hopes or unstated assumptions	Uses real context from the work	Anchored in observed reality and numbers
Accountability	Ownership is diffuse or missing	Named owner exists	Owner, authority, and response path are explicit
Verification	"It looked good" or no check	Some checking performed	Specific checks documented; gaps openly stated
Uncertainty	Overconfident or silent on limits	Mentions some limits	Clearly states what is still unknown or untested
Decision quality	Activity without a clear call	A decision is present	Decision is justified and would change if evidence changed

Passing standard

An artifact is acceptable only if it scores at least "Acceptable" on every row and "Strong" on Accountability and Verification. Anything less is incomplete — revise it.

Checkpoint Rubric (Core Judgment and each Applied Pack)

Before someone moves on, they must demonstrate:

They can apply the relevant filters without being prompted.
They produce artifacts that meet the Core Rubric.
They can explain what would make them change their mind.
They do not treat fluent AI output as finished work.

If any of the four is missing, they stay in the current pack.

How to Use This Section

Learners: Use the rubrics to self-check every artifact before calling it done.
Reviewers / managers: Use the same rubrics so standards stay consistent.
Program owners: Use the implementation timeline and governance model to avoid both chaos and bureaucracy.
Everyone: Treat external frameworks as tools, not as the goal. The goal is better decisions with clear ownership.