Canvas Week Plan

Build a focused weekly action plan by pulling everything due, missing, or upcoming from Canvas into a single prioritised view.

Prerequisites

• Canvas MCP server must be running and connected.
• You must be logged in as a student (or have a student token) — this skill uses student-scoped endpoints.

Steps

1. Load Active Courses

Call get_my_courses to retrieve the student's currently enrolled courses. Note each course ID and name — you will need them for per-course lookups.

2. Fetch Upcoming Assignments

Call get_my_upcoming_assignments to get assignments due within the next 7 days across all courses. Note due dates, point values, and submission types.

3. Check the To-Do List

Call get_todo_items to retrieve any Canvas-generated action items (assignments needing submission, quizzes needing completion). Merge with the upcoming assignments list, deduplicating by assignment ID.

4. Check Upcoming Calendar Events

Call get_upcoming_events to surface any class sessions, office hours, or exam blocks scheduled this week. Include these in the timeline alongside assignment due dates.

5. Review Recent Submissions and Grades

Call get_my_grades to get current grades per course (useful for prioritisation — a course where the grade is borderline deserves more attention this week).

Call get_my_submissions for each course where the student has recent activity to check which assignments are submitted, pending review, or returned with feedback.

For any course whose grade is borderline (near a letter-grade or pass boundary), call project_grade with the course ID and a target — a percentage like 90 or a letter like "A". It returns the minimum uniform percentage needed on all remaining assignments to reach that target, a feasibility flag (achievable / already_secured / impossible), and a plain-language summary. This turns "this course is borderline" into a concrete weekly goal — e.g. "average 84% on the three remaining assignments to hold an A-". project_grade accounts for assignment-group weights and drop rules; it does not factor in late penalties.

6. Check Peer Review Obligations

For each course ID from Step 1, call list_peer_reviews (with the course ID and relevant assignment ID when known) to surface any assigned peer reviews that are due this week but not yet completed.

7. Build the Weekly Plan

Compile everything into a prioritised plan:

1. Overdue / missing — anything already past its due date with no submission
2. Due today or tomorrow — highest urgency
3. Due later this week — plan the effort required
4. Peer reviews due — easy to forget; flag them explicitly
5. Upcoming events — class sessions and exams to block time for

Ask the student if they want to adjust priorities or go deeper on any item.

Output Format

Week Plan — [Student Name]  Week of [Mon date] – [Sun date]

OVERDUE (action required)
• [Course] — [Assignment] — due [date], not submitted
  → Suggest: submit now or talk to instructor

DUE TODAY / TOMORROW
• [Course] — [Assignment] — due [datetime], [points] pts
• [Course] — [Peer review] — due [date] for [peer name]

DUE LATER THIS WEEK
• [Course] — [Assignment] — due [day], ~[estimated effort]
• [Course] — [Quiz] — due [day]

UPCOMING EVENTS
• [Date] [time] — [Event title] ([course])

GRADE SNAPSHOT
• [Course A] — 88%  (on track)
• [Course B] — 71%  (borderline — need ≥84% on remaining work to reach a B)
• [Course C] — 94%  (strong)

Notes

• This skill is fully read-only — it surfaces information but does not submit anything on the student's behalf.
• list_peer_reviews requires a course ID and assignment ID. If the student has many courses, focus on courses where the assignment deadline is within 7 days.
• Grades from get_my_grades reflect the current posted grade and may not include unposted scores. Mention this caveat if a course shows an unexpectedly low or missing grade.
• project_grade uses a uniform-percentage model (the same percentage assumed on every remaining item) and freezes drop rules at their current state — treat its output as a planning target, not a guarantee. It computes one student per call; omit student_id to run it for the logged-in student.