Learning Assessment Rubric

Assessment Rubric

Use this rubric for self-study, onboarding, classroom review, or pull-request review of learning materials.

Area	Beginner	Competent	Advanced	Expert
Workflow execution	Runs the documented commands	Explains each command's purpose	Adapts commands to a new local case	Designs a repeatable workflow for a team
Evidence interpretation	Finds outputs and basic fields	Separates extracted facts from warnings	Explains provenance and uncertainty	Critiques evidence limits and missing review steps
Safety boundary	Repeats that CaeReflex is not a validator	Avoids validation, certification, and safety claims	Corrects unsafe summaries	Designs agent/developer prompts that preserve constraints
CLI and exports	Creates JSON or Markdown outputs	Chooses the right export for a user	Compares full case JSON and agent context	Traces exporter responsibilities in the architecture
REST/OpenAPI	Recognizes health and schema endpoints	Imports a case and retrieves context	Designs a bounded tool-calling workflow	Reviews API exposure, authentication, and workspace risks
Developer architecture	Identifies adapters and services	Explains the service layer's role	Designs a safe adapter change	Plans tests, docs, and safety validation for a new feature

Minimum completion standard

A learner should be able to provide:

• the commands they ran;
• the files or endpoints they inspected;
• one extracted fact;
• one limitation or inspection flag;
• a safe summary that does not claim validation, convergence, mesh adequacy, certification, or design safety.

Instructor review prompts

Ask learners:

1. What evidence came directly from files?
2. What did CaeReflex infer or summarize?
3. What would a qualified engineer still need to verify?
4. Which output would you give to an LLM agent, and why?
5. What statement would be unsafe or unsupported?