agentsclimarketplace

Learnus course copilot

Skill mrcha033/skills/skills/learnus-course-copilot

MrCha Skills — portable Agent Skills for ChatGPT, Codex, and compatible agents

Install
npx -y skills add mrcha033/skills --skill learnus-course-copilot

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • 16 days oldThe repository was created 16 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use the user's authenticated browser session to inspect Yonsei LearnUs courses, collect authorized course materials, identify assignments and deadlines, and produce study-ready structured output. Use for LearnUs course pages, lecture resources, VOD links, assignment tracking, or downloads when credential-based scripts fail or SSO state must remain in the browser.

SKILL.md

2.2 KB, as published. Nobody here has run it

LearnUs Course Copilot

Operate through the user's existing authenticated browser. Never request, store, or pass Yonsei credentials on the command line.

Workflow

  1. Use a browser-control skill and open the exact LearnUs URL in the existing browser session.

  2. Read the current URL and a bounded DOM snapshot.

  3. If the page shows Portal Login, External Login, SSO, or a password field, stop and ask the user to sign in in that browser. Do not automate credentials or copy cookies.

  4. On an authenticated course page, save or pass the page HTML to:

    python3 "$SKILL_DIR/scripts/analyze_learnus_snapshot.py" \
  --html snapshot.html \
  --base-url "https://ys.learnus.org/course/view.php?id=..."

  5. Review the structured result. Use page locators from the fresh DOM snapshot to retrieve only the resources the user asked for.

  6. Ask before triggering a browser download when the browser surface requires confirmation. Do not bypass DRM, access controls, enrollment restrictions, or expiring authorization.

  7. Return the course title, materials, assignments/deadlines, VOD entries, inaccessible items, and the exact pages inspected.

Boundaries

  • Treat public web search as discovery only; it cannot authenticate LearnUs.
  • Do not revive or depend on credential-based learnus-dl.py.
  • Do not infer completion, attendance, or a deadline from a filename alone.
  • Do not redistribute course materials or download an entire course unless the user explicitly asks.
  • Keep signed URLs and tokens out of logs and delivered artifacts. Report a redacted URL or page locator when needed.
  • Mark the result login_required when the snapshot is not authenticated.
  • Follow references/snapshot-contract.md when adding parser support for a new LearnUs page shape.

Run python3 "$SKILL_DIR/scripts/self_test.py" after changing the parser.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.