← Back to the Using Artificial Intelligence outline The Course Maker
← Using Artificial Intelligence outline
Week 12 · Quiz

Week 12 Quiz — Claude Cowork II: Skills, Connectors & Artifacts

Using Artificial Intelligence · AI 101 Fall 2026 · Prof. Quinn Fictional sample

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Covers: skill vs. connector vs. artifact vs. plugin · MCP (open standard, Anthropic) · least-privilege · live artifact (refreshes from connected data) · built-in skills (docx/pptx/xlsx/pdf) · connectors directory
Format: 10 auto-graded items (multiple-choice, multiple-answer, matching, true/false) · 10 points (1 each) · allowed attempts: 1 · No AI on this quiz.

This is the human-readable quiz with its vetted answer key and one-line feedback. The import-ready Classic QTI 1.2 is in F-quiz-week-12-qti.xml (generated by a validated Python script — parses with 10 items, every single-answer item exactly one correct). Reminder: AI is not permitted on quizzes — this checks that you understand the Week 12 ideas.

Questions, key, and feedback

Q1 (Matching). Match each Cowork concept to its correct plain-language definition.

Concept Definition
Skill A bundle of skills + connectors + sub-agents in one installable package
Connector A reusable instruction set (a SKILL.md file) that teaches Claude how to do a specific task
Live artifact A link to an external app, using MCP, that runs with permissions you grant
Plugin A persistent, interactive view that refreshes with current data from your connectors

Correct pairings:
- Skill → A reusable instruction set (a SKILL.md file) that teaches Claude how to do a specific task
- Connector → A link to an external app, using MCP, that runs with permissions you grant
- Live artifact → A persistent, interactive view that refreshes with current data from your connectors
- Plugin → A bundle of skills + connectors + sub-agents in one installable package

Feedback: These four concepts are the week's core — nail the pairings and everything else this week clicks. Memory hook: "Skill = instruction set. Connector = app link. Artifact = live view. Plugin = bundle."

Q2 (MC). What does MCP stand for, and who created it?

  • A. Machine Control Protocol — created by Google
  • B. Model Context Protocol — created by OpenAI
  • C. Model Context Protocol — created by Anthropic as an open standard ✅
  • D. Managed Connector Protocol — a proprietary Claude standard

Feedback: MCP = Model Context Protocol, an open standard created by Anthropic — not a proprietary Claude-only system. "Open standard" means other AI systems can use it too. Sources: verified against support.claude.com and modelcontextprotocol.io.

Q3 (MC). A student installs a "Project Management" package in Cowork. It immediately gives her three custom instruction sets for formatting reports, a Notion connector, a Linear connector, and a sub-agent that triages incoming requests — all in one install. This is an example of a —

  • A. skill
  • B. connector
  • C. live artifact
  • D. plugin ✅

Feedback: A plugin bundles skills + connectors + sub-agents into a single installable package. The tell is "all in one install" — that's a plugin, not any individual component.

Q4 (True / False). A live artifact in Claude Cowork is basically a static file — once Claude generates it, the content doesn't change.

  • True
  • False ✅

Feedback: False. A live artifact is the opposite of static — it refreshes with current data from your connected apps each time you open it. It persists across tasks, keeps version history, and lives in its own tab. (Verified: support.claude.com/en/articles/14729249-use-live-artifacts-in-claude-cowork.)

Q5 (MC). Prof. Quinn asks Claude to "generate a Word document summarizing these project notes." Claude uses a built-in feature to do this. That feature is a —

  • A. built-in skill (the docx skill) ✅
  • B. connector
  • C. live artifact
  • D. plugin

Feedback: Built-in skills include docx, pptx, xlsx, and pdf — document-format skills that let Claude generate real files. No external app is being connected, so this is a skill, not a connector.

Q6 (Multiple answer — select all that apply). Which of the following are TRUE about connectors in Claude Cowork?

  • A. Connectors use MCP (Model Context Protocol) to link Claude to external apps. ✅
  • B. Once you add a Gmail connector, Claude automatically also gains access to your Google Drive.
  • C. A connector inherits your own permissions in the connected service — not more. ✅
  • D. Each connector runs only with the permissions you choose to grant. ✅
  • E. There are exactly 47 connectors in the directory.

Feedback: True: A (MCP), C (inherits your permissions), D (runs with granted permissions). B is false — connectors are scoped separately; Gmail access doesn't automatically include Drive. E is false — the directory is large and grows frequently; never quote a specific count. (Verified: support.claude.com/en/articles/11176164.)

Q7 (MC). A student wants Claude to pull her latest unread emails into a task and draft a briefing. She needs to —

  • A. create a SKILL.md file for Gmail
  • B. connect the Gmail connector in Cowork, granting the appropriate read permissions ✅
  • C. build a live artifact for Gmail
  • D. install a plugin that includes Gmail

Feedback: Pulling external app data requires a connector — specifically, the Gmail connector with read permissions. A skill doesn't pull external data; an artifact is a view built from existing connector data; a plugin is a bundle you install. The task here is the first step: establishing the connection.

Q8 (MC). Which revision to a Cowork setup best applies the principle of least privilege?

  • A. Connect all available Google Workspace connectors at once so Claude never has to ask for access later.
  • B. Connect only the Google Calendar connector with read-only access to the specific calendar needed for the task. ✅
  • C. Grant the Google Drive connector full read and write access so Claude can edit files if needed in the future.
  • D. Skip connectors entirely and paste data manually instead.

Feedback: Least privilege = grant only what the task actually needs, no more. Option B gives exactly what's needed (calendar read) and nothing extra. Option A over-connects; Option C pre-grants write access for a speculative future need; Option D avoids the tool entirely rather than using it safely.

Q9 (MC). A student says: "I built a SKILL.md file that connects Claude to my Notion database and pulls my tasks." What is wrong with this statement?

  • A. Nothing — that's exactly what a SKILL.md file does.
  • B. A SKILL.md file can connect to Notion, but only with enterprise permissions.
  • C. A SKILL.md file is a local instruction set, not a mechanism for connecting to external apps — that requires a connector. ✅
  • D. Skills can only work with built-in document formats; Notion isn't supported.

Feedback: This is the week's central confusion. A skill (SKILL.md) is a local instruction file — it teaches Claude how to do a task. Pulling data from an external app like Notion requires a connector — a vetted MCP server, authorized with your Notion permissions. A skill cannot connect to an external app.

Q10 (MC). You've been using Claude Cowork for three months and have connectors for Gmail, Google Calendar, Slack, and GitHub all set up. You build a new live artifact that shows your open GitHub issues. Which data sources can this artifact access?

  • A. All four connectors you have set up, automatically.
  • B. Only Gmail, since it was set up first.
  • C. Only the connectors you approved specifically when creating or updating this artifact. ✅
  • D. No connectors — artifacts only use Claude's built-in knowledge.

Feedback: A live artifact can only use the connectors you approved when you created or updated it — it doesn't gain automatic access to all your connectors. This is a safety feature: each artifact's data scope is explicit and controlled. (Verified: support.claude.com/en/articles/14729249.)

Answer key (quick reference)

Q Answer Q Answer
1 Skill→instruction set / Connector→app link / Artifact→live refreshing view / Plugin→bundle 6 A, C, D
2 C (MCP, Anthropic open standard) 7 B (connect Gmail connector)
3 D (plugin = bundle) 8 B (least privilege: calendar read-only)
4 False (artifacts refresh) 9 C (skill ≠ connector)
5 A (built-in docx skill) 10 C (connectors approved at creation only)

Blueprint & item-bank note

# Type Concept Objective
1 Matching Four-way distinction: skill / connector / artifact / plugin 5
2 MC MCP = Model Context Protocol, open standard, Anthropic 5
3 MC Scenario: plugin (bundle) 5
4 True/False Artifact is NOT static — it refreshes 5
5 MC Built-in skill (docx) for document generation 5
6 Multiple answer Connector truths (MCP, permissions, not auto-expanded) 5
7 MC Connector required to pull external app data 5
8 MC Least-privilege prompting scenario 5
9 MC Skill ≠ connector (core misconception) 5
10 MC Live artifact uses only its approved connectors 5

All 10 items tagged course=AI101 · week=12 · objective=5. Distractors target classic W12 misconceptions: skill ≠ connector, artifact = static file, plugin = single skill, MCP is proprietary, connector-count specificity, all-connectors-auto-granted.

Quality gate (self-checked)

Canvas placement block

canvas_object    = Quizzes::Quiz
title            = "Week 12 Quiz — Claude Cowork II: Skills, Connectors & Artifacts"
assignment_group = "Quizzes"
points_possible  = 10
grading_type     = points
available_from_offset_days = 0
due_offset_days  = 6
published        = true
allowed_attempts = 1
shuffle_answers  = true
ai_permitted     = false
provenance       = "~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com"
This is the human-readable quiz with its vetted answer key and rationale. The import-ready Classic-QTI version (F-quiz-week-12-qti.xml) ships inside the course's .imscc package — it lands in the Canvas gradebook on import.

~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com