Hire an AI course lesson update reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• Questions per lesson, this month against last, with the repeated question quoted
• Every claim that is now false, with what is true and its source
• A rewrite per flagged lesson with an opener, the smallest working example and a three-level quiz
• Design fixes ranked by impact for lessons that do not need a rewrite
• The lessons due for review next month

How it does the job well

• Base people decisions on job-related evidence, applied the same way to everyone.
• Write in plain, respectful language the person affected could read.
• Keep personal data to what the task needs, and handle it privately.
• Check local employment law and company policy before advising on hiring, pay or exits.

Mistakes it avoids

• Rewriting a whole lesson because one step changed.
• Adding more explanation when the problem is a missing example.
• Treating a quiet lesson as a good lesson.
• Changing three things in one lesson and not knowing which one helped.
• Biased wording in job ads, reviews or interview questions.
• Plans with no owner, timing or way to check progress.

What a good result looks like

An example, not your task: You run a course or training that teaches a tool or process that keeps changing.

• This month. One line with lessons, questions, lessons flagged, rewritten and due next month.
• The question map. Questions per lesson this month against last month, with the repeated question quoted.
• Stale lessons. Each false claim and changed step, with what is true now and where that comes from.
• Rewrites. For each rewritten lesson: an opener, the smallest working example, the body changes and a quiz with answers.
• Design fixes. Ranked fixes for lessons that need changes but not a full rewrite.
• Due next month. Lessons reaching their review date, with the date each was last reviewed.

How the work is checked

• Every flagged lesson has either a question count or a review date behind it.
• Every false claim names what is true now and where that comes from.
• Every rewritten lesson has an opener, an example the learner can check, and a quiz with answers.
• No lesson was rewritten for design fixes alone.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way Andrej Karpathy would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Look at the data. Counts questions per lesson by hand, flags repeats and jumps, and adds lessons past their review date.
• Check what is still true. Tests each flagged lesson against the current docs or code for false claims and drifted steps.
• Check the learning design. Audits each flagged lesson against Universal Design for Learning, with fixes ranked by impact.
• Decide per lesson. Rewrite for false claims or a repeated how question; list fixes for design alone; mark reviewed when nothing is wrong.
• Rewrite the lesson. The smallest example with traced values first, a five-minute opener, the body fixed one thing at a time, and a quiz with answers.
• Set the next review. Updates last reviewed and the next date for every lesson touched.
• Log and hand back. Saves the lesson and question files, then stops. You publish.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id monthly-course-lesson-keeper only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/monthly-course-lesson-keeper/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Your lessons as text, with the date each was last updated
• The month's learner questions, complaints and quiz results, tagged by lesson (advised)
• The current docs or code for what the lessons teach (advised)
• Your review rule, or accept six months
• Last month's keeper file, once it exists

Helpful context (optional)

• Quiz results by lesson, so quiet lessons with low scores are seen
• Release notes for the tool the course teaches

Tools and access

• Claude Code: It can read the lessons folder and the current docs or code, and keep the lesson and question files between months.
• Your course platform: Export learner questions and quiz results by lesson each month.
• Any good chat AI: Working out and tracing a small example is text work any strong chat AI does well with this skill.
• A spreadsheet or calculator: Checking the numbers independently is the quickest test of the trace.
• Any good chat AI: One message is enough.

Functions and deliverables

• Course lesson update reviewer: Find stale lessons and repeated learner questions. Get targeted fixes, examples, lesson openers and recall quizzes.
  • Questions per lesson, this month against last, with the repeated question quoted
  • Every claim that is now false, with what is true and its source
  • A rewrite per flagged lesson with an opener, the smallest working example and a three-level quiz
  • Design fixes ranked by impact for lessons that do not need a rewrite
  • The lessons due for review next month
• Documentation accuracy reviewer: Check documentation claims against the actual code. Get false or unverified claims, source evidence and fixes to make before publishing.
  • A list of false or missing claims
  • Code samples marked as working or broken
  • Places where docs and code disagree
  • Fixes for each issue
  Source: sickn33.
• Accessible lesson auditor: Check your lesson plan for barriers that shut some learners out. Get an accessible lesson audit with fixes ranked by impact, and what to keep.
  • A barrier check across the three UDL principles
  • Fixes ranked by impact
  • Things the lesson already does well
  • Notes on what to leave alone
  Source: Gareth Manning.
• Small working example explainer: Teach how a formula or process works with the smallest worked example. You get traced steps, predictions to try, and a check against a trusted source.
  • The smallest example that still shows the mechanism
  • A trace of every step with named values
  • Two to four predictions with answers
  • A note of what the small version leaves out
• Lesson opener planner: Plan a five-minute lesson opener from today's topic and last lesson. Get a task, what to say, how to read answers and the next move for each.
  • The one idea that links last lesson to today
  • A five-minute opener task that every learner does
  • Two or three sentences to say to start it
  • What ready and stuck answers sound like, and your next move for each
• Recall quiz writer: Turn your lesson notes into a recall quiz with easy, medium and hard questions, model answers and a repeat plan. Ready in one message.
  • The key ideas
  • Nine questions at three levels
  • A model answer for each
  • A repeat plan
• me.md builder: Answer a few questions, or let your AI use what it remembers. Get one private file that tells any AI who you are.
  • One me.md file with up to ten dated sections
  • A list of facts taken from memory, for you to check
  • A short note of what it removed for privacy
  • The two next steps, to save it and set it up
• Andrej Karpathy: Debug an AI project or learn a hard idea using Andrej Karpathy’s recorded approach. Inspect real examples, start small and check each change.
  • Your problem worked on the way Andrej Karpathy would: their questions, their method, their push-back
  • The answer or the work, in Andrej Karpathy's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• You run a course or training that teaches a tool or process that keeps changing.
• The same learner question keeps arriving under the same lesson.
• You want every lesson re-checked on a schedule instead of when someone complains.

Limits to check before starting

• You want one new lesson written from scratch: ask your AI to draft it, or use the Lesson opener planner for just the start of a class.
• You only want to see which quiz topics your class missed: use the Quiz results and reteaching tracker, which reads quiz results after each test.
• Your course teaches a subject that never changes and gets no learner questions: a yearly read-through by you is enough.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Start a chat with the agent. In Claude Code, add it as a skill or save it in .claude/agents, then ask it to run this month.
• Put your lessons in one folder with their last-updated dates. Add the month's learner questions, tagged by lesson.
• Point it at the current docs or code for the tool you teach.
• Read the stale claims and the rewrites. Publish the ones you approve in your course platform.
• Run it again on the first Monday of next month. It does not start itself, so set a reminder or a recurring task. It tracks questions and review dates over time.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/monthly-course-lesson-keeper for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.