Chaz
Community@chaz-clark · 83460
Data Science Professor at BYU-Idaho
Agent Skills by Chaz
Showing 9 vetted skills indexed across 1 GitHub repositories.
chaz-clarkcourse-build
Syncs Canvas LMS course content between local files and the Canvas REST API.
chaz-clarkgrading
Grades Canvas LMS submissions with consensus grading, human review gates, and FERPA de-identification.
chaz-clarkvoicing
Applies an instructor's saved feedback voice profile to student-facing grading comments.
chaz-clarkaccommodations
Applies per-student Canvas LMS accommodations using instructor-supplied de-identified student codes.
chaz-clarktitle-iv
Classifies Canvas LMS student engagement for federal Title IV unofficial-withdrawal compliance audits.
chaz-clarkaudit
Analyzes Canvas course structure and instructional-design quality using read-only audit scripts.
chaz-clarkferpa-deid
De-identifies and re-identifies student submissions and gradebooks for FERPA-safe LLM processing.
chaz-clarkimprove
Manages a git-tracked markdown kanban board for course improvement cards.
chaz-clarkscheduled-peer-review
Configures a GitHub Actions scheduler for deterministic Canvas group peer-review pairing.
Frequently Asked Questions About Chaz
FAQPage SchemaWhat tasks can I automate using Chaz's Canvas skills?▼
You can sync Canvas course content to and from local files, propagate master-to-blueprint-to-section content, run consensus grading with push gates, apply per-student accommodations like quiz time extensions, audit instructional design across 8 frameworks, perform Title IV engagement audits, and manage a course improvement kanban backlog.
Who are these Canvas administration skills designed for?▼
They are built for university instructors and instructional designers, specifically a data science professor context at BYU-Idaho. They serve educators managing multi-section Canvas courses who need grading consistency, FERPA-safe AI processing of student work, disability accommodations, and federal Title IV withdrawal compliance.
How does the FERPA de-identification process work before AI grading?▼
The ferpa-deid skill builds a course-wide de-id master, then de-identifies submissions or gradebooks (docx, pdf, xlsx, jupyter, databricks, text, comments) before anything reaches the LLM. After grading, keyed results are re-identified back to students, with name-leak scanning. Zone-2 files are never read.
What prerequisites and identifiers do the accommodations skills require?▼
The accommodations skill requires the instructor to supply an opaque --deid-code or --user-id for each student; the system never looks up names. It handles SAS/disability accommodations, quiz time extensions, late-work grace, assignment exemptions, group-override recalculation, and submitting on a student's behalf.
How is scheduled peer review pairing set up?▼
The scheduled-peer-review skill sets up and validates an opt-in GitHub Actions scheduler that performs deterministic Canvas group peer-review pairing. It is distinct from the improve skill's kaizen backlog, which explicitly is not continuous integration or GitHub Actions pipeline work.