NTLSN · Crash Course · Claude Code

Claude Code for academics — a crash course

Agentic AI can read, write and run code for you. Four short lessons on using it for academic work without being a developer, then a self-check.

The one thing to remember: you don't need to be a programmer — you need to describe the outcome and review the result. Agentic tools like Claude Code do the coding; you do the judgement.
4 lessons~11 min read1 self-checkGrounded in Claude Code documentation & agentic-AI practice

The lessons

1
What agentic coding is (and isn't)An assistant that does, not just suggests
›

Unlike a chatbot that only answers, an agentic tool can read your files, write code, run it, and check the result — iterating toward a goal you set.

  • You describe the outcome; it plans and executes the steps.
  • It can analyse a spreadsheet, build a page, or wrangle documents.
  • It's powerful and fallible — you stay the decision-maker.
2
What academics can actually do with itReal, useful, non-scary tasks
›

You don't need a research-software project. Small, everyday tasks are where it pays off first.

  • Clean and analyse a dataset, and make a chart.
  • Build a simple in-browser tool, quiz or visual for students.
  • Automate repetitive admin — rename, reformat, summarise files.
  • Draft, restructure and check documents.
3
How to work with it wellSmall steps, verify, version
›

Treat it like a capable junior who needs clear briefs and review. Go in small steps and check as you go.

  • Ask for one thing at a time; review before the next step.
  • Keep versions or backups so you can undo.
  • Make it explain what it did, and test the result yourself.
4
Safety & boundariesData, privacy, and not running blind
›

The power to run code is also a risk. Mind data, permissions and your institution's rules.

  • Never give it confidential or identifiable student/research data without clearance.
  • Don't run commands or code you don't understand on important systems.
  • Check IT, data-governance and AI policies before you start.
Grounded in
  • Claude Code documentation
  • Institutional data-governance & AI policy
◇ Bring it together — from the NTLSN commons
Prompting for Academics courseBuild Your Own Teaching Tools courseFree in-browser tools369 sector videos

Before you try agentic AI for a task — a quick self-check

✓
I describe the outcome and review the result.
✓
I'm starting with a small, everyday task.
✓
I work in small steps and verify each one.
✓
I keep versions so I can undo.
✓
I don't run code I don't understand on important systems.
✓
I've checked data, privacy and AI policy before starting.
Source & attribution. Curated from agentic-AI good practice, including Claude Code documentation, indexed alongside the NTLSN commons. NTLSN is independent and not affiliated with Anthropic or any AI vendor — tools are named as examples. Follow your institution's AI, data and IT policies.
A free NTLSN crash course · see them all at /crash-courses.html. Runs in your browser; nothing is stored or sent.  ·  ← Back to NTLSN