Claude 101

Claude 101: Complete Study Guide for MCQ Exams

Core Concepts

What is Claude?

Definition: Claude is an AI assistant designed to be a thinking partner—more than a chatbot, it’s capable of conversational and text processing tasks while maintaining reliability and predictability.

Guiding Principles:

• Helpful: Designed to assist with complex problems
• Harmless: Built to avoid toxic or discriminatory outputs
• Honest: Transparent and truthful in interactions
• Constitutional AI Approach: Trained to align with human values

Key Differentiators: | Feature | Claude | Simple Chatbot | |———|——–|—| | Scope | Wide variety of complex tasks | Simple Q&A | | Collaboration | Steerable and collaborative | Limited customization | | Reliability | High degree of predictability | Variable | | Tasks | Summarization, research, writing, coding, reasoning | Basic responses |

Claude Capabilities

Core Capabilities

1. Writing and Content Creation

• Collaborative writing on emails, reports, social media
• Iterative refinement until your voice comes through
• Takes direction on personality and tone

2. Research and Analysis

• Explores research angles and compiles findings
• Analyzes data to surface insights
• Context Window: 200K+ tokens standard (500+ pages)
• Extended Context: Up to 1M tokens available on Pro, Max, Team, Enterprise

3. Coding Assistance

• Claude Opus 4.7 is the best coding model
• Helps write, debug, and explain code across languages
• Real-world problem solving

4. Problem-Solving and Reasoning

• Complex cognitive tasks and strategic thinking
• Mathematical problems
• Extended Thinking: Step-by-step reasoning for complex analysis

5. Learning and Development

• Adapts to your learning style and pace
• Learning Mode: Guides reasoning rather than providing answers
• Develops critical thinking skills

Access Methods

Ways to Access Claude

Platform	Best For	Features
Claude.ai (Web)	General conversations, writing, research, analysis	Primary interface; syncs across devices
Claude Desktop App	Integration with local tools; Chat, Cowork, Code modes	Native to machine; screenshots, dictation (Mac)
Claude Mobile Apps	On-the-go access	Sync conversations across devices
Claude Code	Development workflows	Terminal, IDE, browser integration
Claude in Slack	Team collaboration	Direct channel/thread access; @Claude mentions
Claude for Excel	Spreadsheet analysis and modeling	Sidebar in Microsoft Excel
Claude for PowerPoint	Presentation creation and editing	Sidebar in Microsoft PowerPoint
Claude for Chrome	Web research and browser automation	Chrome extension sidebar (Research Preview)

Plan Availability:

• Free, Pro, Max: Claude.ai + mobile apps
• Pro, Max, Team, Enterprise: Code, Cowork, Research, Skills
• Team & Enterprise: Collaboration features, Enterprise Search
• Chrome extension: Research preview on Pro, Max, Team, Enterprise

Key Features

1. Projects

Definition: Self-contained workspaces with persistent memory, knowledge bases, custom instructions, and chat histories.

Components:

• Knowledge Base: Upload documents (PDF, DOCX, CSV, TXT, images) that Claude references
• Project Instructions: Custom behavior guidelines applied to every conversation
• Team Collaboration: Share with specific permissions (View, Edit, Owner)

When to Use:

• Ongoing work with reference materials
• Consistent requirements for responses
• Team collaboration needs
• Knowledge that needs to persist across conversations

Scaling: When knowledge approaches context limits, Claude enables RAG (Retrieval Augmented Generation) mode—expands capacity 10x while maintaining quality.

Permission Levels:

• Can view: Read-only access with discussion rights
• Can edit: Full collaboration power; modify instructions and knowledge
• Owner: Full control including member management

2. Artifacts

Definition: Standalone, interactive outputs Claude creates in a dedicated window alongside conversations.

Auto-Created When:

• Content is significant and self-contained (typically 15+ lines)
• Represents complex content standing on its own
• Meant to be edited, iterated, or reused
• Will be referenced or used later

Types of Artifacts:

• Documents (Markdown, Word, PDF, PowerPoint, Excel)
• Code snippets (Any programming language)
• HTML pages (Complete web pages with CSS/JavaScript)
• SVG images (Scalable vector graphics)
• Mermaid diagrams (Flowcharts, sequence, Gantt, org charts)
• React components (Interactive UI with functionality)

Actions Available:

• View preview or code
• Copy content
• Download to computer
• Share within organization (Claude for Work)
• Publish publicly (Free, Pro, Max users)
• Remix (Others can modify in their conversations)

Publishing: Anyone with link can access; not indexed by search engines

3. Skills

Definition: Instruction packages that teach Claude specialized workflows; expertise packages encoding repeatable processes.

Types:

• Anthropic Skills: Built and maintained by Anthropic (Excel, Word, PowerPoint, PDF creation)
• Custom Skills: Created by you or your organization for domain-specific tasks

Availability: Pro, Max, Team, Enterprise (Feature preview); requires Code execution enabled

How They Work:

• Claude automatically invokes relevant skills
• Procedural machines encoding HOW to do something
• Consistent methodology applied every time

Creating Custom Skills:

• Start conversation with desired workflow
• Claude asks clarifying questions
• Upload reference materials/examples
• Claude generates properly structured skill file
• Automatically applied to relevant tasks

Skills vs. Projects:

Aspect	Skills	Projects
Purpose	Define processes Claude executes	Store knowledge Claude references
Best For	Repeatable workflows, multi-step tasks	Long-term context, reference materials
Persistence	Applied when skill is invoked	Available across all chats in project
Example	Brand guidelines, blog drafting process	Customer hub, research buddy

4. Connectors

Definition: Tools that give Claude access to your data and applications; powered by Model Context Protocol (MCP)—universal standard like “USB-C for AI.”

Types:

• Web Connectors: Cloud services (Gmail, Notion, Slack, Asana, Linear, Stripe, etc.)
• Desktop Extensions: Local tools accessed via Claude Desktop app

How They Work:

• Read information from connected tools
• Perform actions on your behalf
• Search files, retrieve documents, update records
• Execute tasks across applications

Setting Up Web Connectors:

1. Navigate to claude.ai/directory
2. Click “Connect”
3. Authenticate with service login
4. Grant permissions
5. Test connection

Common Use Cases:

• Project Management: Asana, Linear, Jira → “What are my highest priority tasks?”
• Communication: Slack, Gmail → “Find email thread about vendor contract”
• Documentation: Notion, Google Drive, Confluence → “Search for brand voice guidelines”
• Business Tools: Stripe, PayPal, Salesforce → “Show revenue trends”

Security:

• Scoped access (specific permissions)
• Claude sees only what you see
• Revocable anytime through settings

5. Enterprise Search

Definition: Dedicated “Ask {Organization Name}” option in sidebar; pre-built project for entire organization’s knowledge base.

Purpose: Find and synthesize knowledge buried across company’s tools and data sources

Data Sources: SharePoint, Slack, Gmail, Google Drive, Microsoft Teams, Email

Use Cases:

• Getting Up to Speed: “What happened yesterday?” “Summarize key updates”
• Policy Questions: “What’s our remote work policy?”
• Research: “What are main reasons customers choose competitors?”
• Onboarding: “How does our authentication system work?”
• Performance Tracking: “Summarize team contributions to X initiative”

Setup (2-Step Process):

1. Admin: Configure connectors (Documents required, Chat required, Email optional)
2. Users: Authenticate with personal accounts

Safety: Only shows data user already has access to; conversations remain private

6. Research Mode

Definition: Advanced feature transforming Claude into systematic investigator; conducts multiple searches building on each other.

Key Characteristics:

• Agentic Process: Claude autonomously determines next steps based on findings
• Extended Thinking: Automatically enabled; allows step-by-step planning
• Timeline: 5-15 minutes for most reports; up to 45 minutes for complex investigations
• Citations: All claims link to sources for easy verification

How It Works (4 Steps):

1. Claude plans approach (Extended thinking activated)
2. Conducts multiple linked searches
3. Synthesizes findings from web and connected integrations
4. Provides citations for every claim

When to Use Research:

• Comprehensive reports from multiple sources
• In-depth analysis requiring hours of manual work
• Comparative analysis (competitors, vendors)
• Thorough investigations with verifiable claims

When NOT to Use Research:

• Quick facts (stock price, company address)
• Single-source answers
• Speed more important than comprehensiveness
• Internal-only questions (use Enterprise Search instead)
• Deep reasoning without external info (use Extended Thinking)

Prompt Tips:

• Be specific about goals
• Specify desired structure/sections
• Include relevant constraints (budget, timeline, geography)
• Ask Claude to help refine prompt beforehand

Feature Comparisons

Claude Desktop App Modes

Feature	Chat	Cowork	Code
Optimized For	Quick exchanges, exploration, iteration	Complex sustained work, research, analysis	Building software, development workflows
Key Strength	Immediate availability, low friction	Multi-source synthesis, scheduled tasks	Direct codebase integration, git tracking
Interface	Compact overlay possible (Mac)	Sidebar with task tracking	Full development environment
Tools Access	Connectors, Skills	Connectors, Plugins, Subagents	Full filesystem, terminal, dev tools
Work Scope	Single conversation	Multiple parallel tasks	Multiple projects/sessions
File Access	Desktop connectors	Local folder access (contained)	Full filesystem access (local/remote)
Code Execution	No	Yes (with connectors)	Yes (primary focus)
Collaboration	Individual	Team via Projects	Team via remote/local sessions

Unique Features by Mode:

• Chat: Quick entry (Mac double-tap), screenshots, dictation, instant overlay
• Cowork: Scheduled tasks, subagents, folder access, computer use, plugins
• Code: Ask/Code/Plan modes, visual diffs, git integration, local and remote environments

Projects vs. Skills vs. Artifacts

Aspect	Projects	Skills	Artifacts
What Is It	Workspace	Process template	Output/document
Persistence	Persistent across chats	Applied when invoked	Single instance
Purpose	Store knowledge	Define process	Display result
Team Use	Share with permissions	Organization-wide	Publish publicly
Examples	Brand hub, research buddy	Blog template, compliance checklist	Landing page, flowchart
Auto-Created	Manual upload	Manual or conversation	Automatic when criteria met

Prompting: Web Search vs. Research vs. Extended Thinking

Mode	Best For	Timeline	Output	Requires
Web Search	Quick facts, current info	Instant	Direct answers with citations	Web search enabled
Research	Comprehensive multi-source analysis	5-45 minutes	Detailed report with citations	Research feature enabled
Extended Thinking	Deep reasoning, complex logic	Varies	Thoughtful analysis	May take longer

Choosing Between Them:

• Need current info from web? → Web Search
• Need thorough investigation? → Research
• Need deep reasoning on internal problem? → Extended Thinking
• Need all three? → Research (includes Extended Thinking + web)

AI Fluency Framework

4D Framework for AI Fluency

Definition: Four core competencies enabling effective collaboration with AI tools.

1. Delegation

• Deciding what work humans should do vs. AI
• Understanding goals and AI capabilities
• Strategic task distribution

2. Description

• Effectively communicating with AI systems
• Clearly defining outputs
• Guiding processes and specifying behaviors
• Prompt Framework: Setting stage → Defining task → Specifying rules

3. Discernment

• Evaluating AI outputs critically
• Assessing quality, accuracy, appropriateness
• Determining improvement areas
• Evals: Testing Claude on recurring tasks (5-10 examples)

4. Diligence

• Using AI responsibly and ethically
• Making thoughtful choices about systems
• Maintaining transparency
• Taking accountability for AI-assisted work

Effective Prompt Structure

Three Elements:

1. Setting the Stage
   • Your role and objectives
   • Context about your work
   • Why this matters
2. Defining the Task
   • Specific action (write, analyze, build, etc.)
   • Relevant details
   • What success looks like
3. Specifying Rules
   • Style and tone
   • Format requirements
   • Examples of desired output

Important Definitions

Term	Definition
Artifacts	Standalone interactive outputs Claude creates in dedicated window (code, documents, diagrams, etc.)
Constitutional AI	Training approach aligning Claude with human values; focused on being helpful, harmless, and honest
Context Window	Amount of text Claude can process; standard 200K+ tokens (~500 pages); extended up to 1M tokens
Connectors	Tools linking Claude to applications (Gmail, Notion, Slack) via Model Context Protocol
Custom Skills	User/organization-created instruction packages for domain-specific workflows
Delegation	AI Fluency competency: deciding what work should be human vs. AI
Description	AI Fluency competency: effectively communicating with AI systems
Discernment	AI Fluency competency: evaluating AI outputs critically
Diligence	AI Fluency competency: using AI responsibly and ethically
Enterprise Search	Dedicated organizational knowledge base search (“Ask {Org}”) across all connected tools
Evals	Systematic testing of Claude’s performance on specific task types
Extended Thinking	Step-by-step reasoning mode for deep analysis; automatically enabled with Research
MCP (Model Context Protocol)	Universal standard for AI tool connections; like “USB-C for AI”
Projects	Self-contained workspaces with knowledge base, custom instructions, team collaboration
RAG (Retrieval Augmented Generation)	Mode enabling Claude to search project knowledge intelligently; expands capacity 10x
Research Mode	Agentic feature conducting multiple linked searches; synthesizes findings with citations
Skills	Instruction packages encoding repeatable processes and methodologies
Thinking Partner	Claude’s primary role; collaborative assistant for complex problem-solving

Critical Theory for MCQ Success

When to Use Each Feature

Use Projects When:

• You have reference materials to reuse repeatedly
• You want consistent behavior across conversations
• You need team collaboration
• Knowledge must persist

Use Skills When:

• You have repeatable workflows
• You want consistent methodology
• Multiple people need same process

Use Artifacts When:

• Creating substantial standalone content (15+ lines)
• Content will be edited or iterated
• Need to share or download result

Use Connectors When:

• Need to access external tools/data
• Want to automate data gathering
• Data exists in external systems

Use Research When:

• Need comprehensive multi-source analysis
• Investigation would take hours manually
• Citations/verification needed
• ✗ Just want quick fact (use web search)

Use Enterprise Search When:

• Question specific to organization
• Need internal knowledge synthesis
• Drawing from multiple internal sources

Common MCQ Patterns

Definition Questions

Format: “What is X?” Strategy: Match exact definition from course material

Capability Questions

Format: “Claude can help with…” Strategy: Know the 5 core capabilities

Feature Purpose Questions

Format: “When should you use X?” Strategy: Know specific use cases for each feature

Comparison Questions

Format: “What’s the difference between X and Y?” Strategy: Use the comparison tables above

Best Practice Questions

Format: “Which is the best way to…” Strategy: Know the 4D Framework principles

Context/Constraint Questions

Format: “Given scenario X, which feature would…” Strategy: Match scenario to appropriate feature

Pro Tips for Exam Success

1. Remember The Rule of 15: Artifacts auto-create for substantial content (typically 15+ lines)

2. Context Window Facts:
   • Standard: 200K+ tokens (~500 pages)
   • Extended: 1M tokens on higher plans
3. Plan Availability Pattern: Higher tiers get more features
   • Free: Claude.ai only
   • Pro: +Code, Cowork, Research, Skills
   • Max: Same as Pro
   • Team/Enterprise: +Collaboration, Enterprise Search
4. Feature Relationships:
   • Projects provide WHAT (knowledge)
   • Skills provide HOW (process)
   • Artifacts provide OUTPUT (result)
5. Permission Levels (Projects):
   • View < Edit < Owner (always remember this order)
6. Three Elements of Good Prompts:
   • Setting stage + Defining task + Specifying rules = Success
7. When Confused About Feature:
   • If PERSISTENT across conversations → Project
   • If METHODOLOGY focused → Skill
   • If EXTERNAL data needed → Connector
   • If INTERNAL org knowledge → Enterprise Search
   • If MULTI-SOURCE investigation → Research

Quick Reference Checklist

Core Concepts to Know:

• What Claude is (thinking partner, not just chatbot)
• Three guiding principles (helpful, harmless, honest)
• Constitutional AI definition
• 5 core capabilities

Features to Know:

• Projects: definition, components, permissions, RAG
• Artifacts: types, auto-creation criteria
• Skills: types, custom vs. Anthropic, Skills vs. Projects
• Connectors: MCP, web vs. desktop, setup steps
• Enterprise Search: purpose, setup, data sources
• Research: agentic process, when to use, timeline

Access Methods:

• All 8 ways to access Claude
• Plan availability for each

AI Fluency:

• 4D Framework: Delegation, Description, Discernment, Diligence
• 3 prompt elements
• What evals are

Comparisons:

• Chat vs. Cowork vs. Code
• Projects vs. Skills vs. Artifacts
• Web Search vs. Research vs. Extended Thinking

Study Strategy

For Quick Review (5 minutes):

1. Review definitions section
2. Check Feature Comparisons tables
3. Verify When to Use patterns

For Deep Study (30 minutes):

1. Read each feature section
2. Work through example scenarios
3. Practice matching scenarios to features

For Exam Practice:

1. Create flashcards from definitions
2. Practice comparison questions
3. Do scenario matching exercises

---

I have revised through the note and have gotten the certification.

View Certificate - Click to view or download PDF