CS-215 AI Literacy
Week 2 homework · completion check

Agent Setup Readiness

Prepare one coding agent for later course practice—or document the blocker and complete the same reasoning task from a supplied transcript. This checks readiness and safe boundaries, not purchasing power or prior terminal experience.

CHOOSE ONE ROUTEInstalled, blocked + transcript, or supervised lab machine.
USE SYNTHETIC DATAOne supplied file; no personal or course work.
SUBMIT FIVE ITEMSA short LMS response; no screenshots or logs.
Boundary first

This is setup practice—not permission for unsupervised coursework.

Week 2 teaches what an agent can observe and request. Do not use the agent for substantive course work yet; later weeks add stronger review, privacy, writing, testing, and deployment controls.

Use

The supplied synthetic file in a new disposable folder. Ask only the bounded read-and-explain question below.

Pause

Before any permission request, read what the tool wants to access or run. Keep the task read-only.

Do not use

No coursework draft, personal file, secret, credential, private repository, institutional record, or real client data.

Stop rule: if the tool requests file edits, shell commands, web access, broader folders, payment, or information you do not understand, deny the request and use the blocked route.
Step 1

Choose the route your current access supports.

All three routes earn the same completion credit. Choose a product only if its current account and device requirements work for you.

Installed route

Step-by-step help for Claude Code, including what to do when sign-in refuses your account: Set Up Claude Code mini workshop.

  1. Check the current official documentation and account requirements.
  2. Install only from the official page; ask for help rather than bypassing device or organization controls.
  3. Open a terminal in a new disposable folder and verify that the tool launches.
  4. Run the supplied read-only practice below. Deny any request outside its boundary.

Codex CLI

OpenAI’s current guide covers installation, sign-in, project-directory use, permissions, and review. Start with codex; if a version command is available, record its output without account details.

Official Codex CLI guide ↗
Step 2

Run one bounded read-only task.

This is deliberately small. The goal is to observe the agent’s input, instruction boundary, evidence, and stopping behavior.

01

Create a disposable folder.

Name it week02-agent-practice. Do not place it inside a real course, work, or personal repository.

03

Launch the agent from only that folder.

Confirm the displayed working directory is the disposable folder before prompting. If you cannot confirm it, stop and use Route B.

04

Give this exact bounded prompt.

Read only week-02-agent-setup-practice.md. Do not edit files, run commands, browse the web, or take any external action. Identify which candidate path follows the stated naming rule, cite the exact rule and candidate, then stop.
05

Judge the response; do not merely trust it.

The supported answer is Candidate A, week-03.html, because the file says weekly hub filenames use two digits after week-. Record whether the agent used that evidence and respected the stop rule.

Step 3 · Route B

No install? Analyze the same interaction.

This supplied transcript measures the same ideas—scope, evidence, and stopping—without pretending you ran the tool.

STUDENT PROMPT
Read only week-02-agent-setup-practice.md. Do not edit files, run commands, browse the web, or take any external action. Identify which candidate path follows the stated naming rule, cite the exact rule and candidate, then stop.
SUPPLIED AGENT RESPONSE
Candidate A, week-03.html, follows the rule “Weekly hub filenames use two digits after week-.” Candidate B uses one digit. I made no changes and took no external action.
OBSERVATIONWhat file evidence supports the answer?
BOUNDARYWhich requested limits did the response say it followed?
LIMITATIONWhat can this transcript not prove about a live installation?
Step 4

Submit one short readiness record.

Use the LMS item named Week 2 Agent Setup Readiness. The LMS supplies the exact due date and any resubmission policy.

Five required items

  1. Route: A installed, B blocked + transcript, or C supervised.
  2. Tool or blocker: Claude Code/Codex plus version or successful launch; otherwise OS, official guide, stopped stage, and concise blocker.
  3. Practice conclusion: which candidate path follows the rule, with the supporting rule.
  4. Boundary check: one permission or safety decision you observed.
  5. Next step: ready for later practice, or the support you need.

Suggested length: 150–250 words or an instructor-approved accessible equivalent.

Completion standard: all three routes can earn full completion. A successful installation receives no advantage over a careful blocker diagnosis plus transcript analysis.