// live AI grading, streaming feedback, workflow practice, credentials, proof exports

$ map --grading="coverage"

AIducation now exposes which learning surfaces are live AI graded, streamed, hybrid, heuristic practice, or evidence rollups. That keeps the platform honest: demos use the universal grader, courses produce proof, and unresolved grading gaps are visible.

[VIEW_COVERAGE_API][LIVE_AI_SURFACES][GRADING_GOVERNANCE]

Coverage summary

8
Surfaces
6
Live AI
385
Graded objects
16
Visible gaps
[!] Tool missions and workflow steps still have local practice scoring in places; high-stakes proof routes through graded evidence.

// Coverage_loop

The coverage map separates real grading, practice scoring, and evidence rollups so implementation priorities stay concrete.

Submit

Learners submit role demos, scenario responses, workflow outputs, tool artifacts, or exit evidence.

Grade

Live AI grading handles demos and scenarios while hybrid surfaces create practice and artifact evidence.

Govern

Guardrails, cost controls, output validation, and tenant scope protect high-impact grading paths.

Prove

Validated results feed manager reports, credentials, proof artifacts, and enterprise exports.

// Grading_surfaces

Each surface lists its endpoint, implementation file, evidence produced, readiness handoff, and remaining gaps.

View role demo coverage
/api/grading/practice

Role demo grading

live ai

The public role picker and try-demo flow submit learner responses to the practice grading bridge, which delegates to universal AI grading with role-specific demo targets.

15
objects
15
roles
2
gaps
Evidence
  • - Overall score
  • - Rubric dimension scores
  • - Strengths and improvement areas
  • - Manager-ready coaching summary
Remaining gaps
  • - Persist anonymous demo attempts only when a tracked demo session exists
  • - Expose more role-specific sample answers for evaluator calibration
/api/grading

Authenticated scenario grading

live ai

Course and dashboard scenario practice posts to the guarded universal grader and can persist identity-bound results.

15
objects
15
roles
2
gaps
Evidence
  • - Scenario attempt result
  • - Rubric feedback
  • - Skill and failure-mode evidence when a rubric contract exists
  • - Learner progress callback
Remaining gaps
  • - Keep expanding tenant-scoped custom scenarios beyond platform fallbacks
  • - Add more per-role benchmark calibration examples
/api/grading/stream

Streaming grading experience

streaming ai

The streaming grading hook uses the same universal grading service with progress events for richer learner feedback.

7
objects
15
roles
2
gaps
Evidence
  • - Progress events
  • - Final grading result
  • - Dimension scores
  • - Skill atom and failure-mode detail when available
Remaining gaps
  • - Use streaming mode in more public practice surfaces
  • - Expose persisted target metadata in learner-facing result cards
/api/grading/practice

Workflow lab grading

live ai

Workflow labs can route final submissions through practice grading before completed artifacts move into portfolio and manager evidence.

15
objects
15
roles
2
gaps
Evidence
  • - Prompt score history
  • - Accepted step outputs
  • - Workflow completion summary
  • - Average prompt score
Remaining gaps
  • - Persist richer step-level rubric evidence outside the local player state
  • - Move every workflow player completion action from local scores to practice grading result ids
/api/grading/practice

Tool mission AI grading

live ai

High-stakes tool mission submissions use practice grading before they can count as readiness or credential evidence.

33
objects
15
roles
2
gaps
Evidence
  • - Mission completion
  • - XP and attempt score
  • - Tool-specific output
  • - Verification challenge evidence
Remaining gaps
  • - Keep low-stakes tutorial steps clearly separate from readiness evidence
  • - Replace visual mock outputs in creative simulators with saved artifact evidence
/api/grading/practice

Exit assessment evidence rollup

live ai

Exit assessments use practice grading for final scenario submissions, then combine baseline, diagnosis, capstone, portfolio, and manager-review evidence into a credential decision.

15
objects
15
roles
2
gaps
Evidence
  • - Readiness delta
  • - Scenario queue
  • - Blocking pass criteria
  • - Credential handoff
Remaining gaps
  • - Persist formal exit attempts as first-class DB records
  • - Add reviewer sign-off workflow for regulated roles
/api/credentials/catalog

Credential evidence verification

evidence rollup

Credentials verify readiness by linking completed modules, graded scenarios, rubric evidence, workflow artifacts, and manager proof.

60
objects
15
roles
2
gaps
Evidence
  • - Credential evidence artifacts
  • - Proof requirements
  • - Verification path
  • - Linked primitive counts
Remaining gaps
  • - Require signed evidence hashes for issued production credentials
  • - Add renewal attempts for expiring skill evidence
/api/proof-artifacts

Course proof artifact coverage

evidence rollup

Proof artifacts make course completion usable by portfolios, manager reports, credentials, LMS, HRIS, audits, and company academies.

225
objects
15
roles
2
gaps
Evidence
  • - Course output artifacts
  • - Simulation result references
  • - Manager report records
  • - Export-safe evidence records
Remaining gaps
  • - Add production attachment storage for uploaded artifacts
  • - Add artifact-level approval status and reviewer identity