Lab 2 — Codebase Detective
START HERE · YOUR LAB FILES
Open these resources
Each card opens the supplied file. Keep these alongside the lab; open the same paths in your downloaded course folder when using your AI.
- 1AGENTS.mdCourse guide · read before starting
AGENTS.md - 2allocation-requirements.mdSupplied evidence · read this source
fixtures/allocation-requirements.md - 3github-issues.jsonSupplied evidence · read this source
fixtures/github-issues.json - 4ai-task-brief.mdTemplate · copy into your own work
templates/ai-task-brief.md
Underlined filenames throughout the steps are clickable. A file you create yourself links to saving instructions, not a nonexistent download.
On this page
SESSION 02 / 35-minute practice lab / within a 2-hour session
Use an AI assistant to investigate a code path while keeping scope and verification in your hands.
A bounded brief, three verified code references, a tool result and a product question.
Before you start · recall, then check
Without opening yesterday’s notes, explain the difference between a plausible explanation and a finding. Then check your explanation against one saved source.
Use about two minutes of the existing arrival/discussion time. Try in your own words, then check the source with a partner or coach. This is ungraded practice, not a new submission.
Need a hand? A supported route
Ask for one function or test at a time. Open every cited location yourself. Use the supplied test before considering a new one; investigation does not authorize a permission change or message send.
Time: 35 minutes
Team: driver and verifier
Quick start
First action: Read AGENTS.md, requirement R1, and issue G102, then trace the Send inquiry update request.
Finish with: A task brief, three verified code references, a check result and an open question.
Step-by-step help for Session 02 — smaller steps, checkpoints and troubleshooting. Saving and asking for help.
Open these three resources first
| Open online | What to read | File in your downloaded course folder |
|---|---|---|
| AGENTS.md — instructions for the AI | Read the course boundaries; you do not need to edit this file. | AGENTS.md, beside package.json |
| Requirement R1 — who may send | Find the line beginning R1:. | fixtures/allocation-requirements.md |
| Issue G102 — Concierge send-update returns 403 | Find G102 in the issue list. It is not “G2.” | fixtures/github-issues.json |
R1 and G102 are labels inside those files, not filenames. Read them online using these links, then open the same files in your course folder when working with your AI.
Mac / Linux / Windows
Submission: prepare the deliverables below using the public-safe packet and review workflow. One PR per lab; private review is equally available. Never publish grades or personal information.
Tell Codex whether you use macOS zsh, Linux bash, Windows PowerShell or Command Prompt before requesting commands. Open the extracted repository in your editor, preserve filename case, and run npm test (npm.cmd test if PowerShell blocks npm). Do not use another learner's absolute paths.
See the platform guide and included local lab track.
Scenario
The training CRM’s “Send inquiry update” button appears to do nothing for staff users. You are asked to investigate the report, not to rush a fix.
Mission
Use Codex and repository tools to map the relevant request path and create a focused investigation plan. You may read and run approved tests; do not edit production-like settings or send a message.
Steps
- Read AGENTS.md, the report in this handout, R1 and issue G102. The staff report and concierge issue are different claims; investigate each without assuming either demands a permissions change.
- Complete an AI task brief asking Codex to locate the relevant code and tests.
- Ask Codex for a plan before any proposed code change.
- Independently verify the files/functions it cites.
- Run one approved existing test or static check.
- Write your investigation result: likely path, evidence, open question, and recommended next step.
Deliverables
- AI task brief
- repository map with at least three verified paths/references
- one tool/test result
- one question that requires a product owner or engineer rather than a guess
Success rubric
A successful team is able to say “we do not yet know” precisely. A speculative root cause is not a finding.
Stretch
Have Codex propose a small test addition. Review the diff for scope, selector/fixture stability, and whether the expected behavior is actually specified.
Product Studio homework — Kick off your own repository
Finish the approved-proposal PR, three initial issues, README and project instructions. Confirm your client/model/host and spending ceiling. Build only the first small approved slice; arrange human feedback on the kickoff PR.
Budget two to three hours, including evidence and review preparation; ask for coaching or scope adjustment if blocked. Record progress using the milestone template in your own product repository, not the practice packet. See the complete project roadmap, GitHub workflow and small-model guide. The practice lab above remains separate from your individual product.
Leave with a next step
Use the existing closing/debrief time: explain one claim you verified, one limit and your next action. Reuse the lab’s artifacts; this adds no separate worksheet or graded test.
Practice evidence uses the lab submission workflow. Product Studio work belongs in your own product repository. These are different destinations.
Product Studio roadmap · Sharing and review guide
Opening the next session does not mark this one complete. Progress tracking is optional and does not award a credential.