Skip to content

Learn Tracks Reference Guides Saved

ع

Learn Tracks Reference Guides Saved

learn → Level 5: Trust & Safety

L05 · learn

Trust & Safety

Read the prompt, judge the blast, pick the mode, and know what brings you back when something goes wrong.

learn 9 lessons

✦

preview · level 5

See it before you read it.

Read four real permission prompts and decide: allow or deny? Learn to read blast radius in a minute.

open the preview →

01Claude proposes, you approveread✓→
02Anatomy of a permission promptread✓→
03Blast radiusread✓→
04One-way doorsread✓→
05The permission modesread✓→
06Allowlist the boring stuffread✓→
07Trusting a projectread✓→
08Git is your undoread✓→
09The pre-flight checklistread✓→

◆ Independent — not affiliated with, or endorsed by, Anthropic. © 2026 Claude for Humans Course updates

ESC

Pages

Claude Code, by what you can do with it.

Pick your team. Every playbook is a worked example with the exact prompts to copy — plus a few starter recipes to get going fast.

In-depth, plain-language guides to getting the most out of Claude in 2026 — power-user Claude Code workflows, what Claude can do…

Guides

9 Claude Code workflows that actually save time

What makes Claude Code fast isn't what you ask for — it's how you set up the work. These are the nine habits that separate a frus…

Claude Code Desktop vs the terminal (CLI): which should you use?

They're the same Claude underneath — only the interface differs. Claude Code Desktop is a normal app you point and click; the CLI…

Claude Code vs ChatGPT for business work (2026)

ChatGPT is a chat window you copy answers out of; Claude Code is an agent that works inside your actual files and folders and sho…

Claude Code vs Cursor vs GitHub Copilot (2026)

They sit at different points on one spectrum — Copilot is autocomplete inside your editor, Cursor is an AI-first editor you steer…

Context window vs Memory in Claude

The context window is what Claude sees right now — this session's messages, files, and outputs. Memory is what it remembers acros…

Dynamic workflows: when Claude builds its own harness

Claude Code can now write a custom orchestration script on the fly — spawning and coordinating its own subagents for the task in…

Features

Hand a self-contained task to a second Claude with its own fresh context, so your main conversation stays focused.

Workflows & agent teams

One request fans out into many Claudes working in parallel — each on a piece — then their results are gathered back together.

Agent view & background work

One screen for every Claude you have running — start jobs, send them to the background, and get pinged when one needs you.

Monitor (watch & react)

Claude watches a running command's output live and reacts the moment something happens — an error, a passing test, a new log line.

A live side panel where Claude builds something you can see, use, and tweak — a document, a chart, a working mini-app — right nex…

A workspace that holds related chats plus shared knowledge — files, a style guide, custom instructions — so every chat in it star…

For your team

You already know how to code. This is the part that's new — handing real work to an agent that reads the whole repo, writes the t…

Turn a folder of notes, briefs, and past campaigns into drafts, calendars, and performance analysis — in your brand voice, all fr…

Research accounts, personalize outreach at scale, prep for every call, and keep your pipeline notes clean — all from the Claude D…

Wrangle spreadsheets, reconcile exports, run the recurring ops summary, and turn raw numbers into clear answers — with the formul…

Draft job posts, balance screen notes, rewrite policies, and build onboarding docs — consistent, inclusive, and fast — all from t…

Data & Analysts

For business analysts and anyone who turns data into decisions — profile a dataset, write and explain a query, turn a CSV into a…

Playbooks

Map every stakeholder before you write a single requirement

Turn a messy list of names into a stakeholder map — who cares, what they need, their influence and interest, and a RACI per key d…

Build the requirements doc every downstream playbook inherits

Turn scattered stakeholder interview notes into one requirements doc — background, in-scope/out-of-scope, functional and non-func…

Turn interview notes into a clear as-is process map

Turn scattered interview notes, screenshots, and a walkthrough into a clear as-is process map — step by step, who does what, with…

Find the gaps between today and the target state

Compare the current state against the future state in your requirements doc, and get every gap named with its root cause — not ju…

Turn scattered concerns into a SWOT and a living risk register

Turn a pile of half-formed worries and notes about an initiative into a structured SWOT and a risk register with likelihood, impa…

Compare solution options and land on one

Score 2-4 candidate solutions against the same criteria — cost, effort, risk, time-to-value — and get a defensible recommendation…

Recipes

Build a basic dashboard

Take a spreadsheet of numbers and turn it into a simple web page with charts you can open in your browser — no coding, no termina…

Clean up and analyze an Excel file

Hand Claude a messy spreadsheet and get back a clean version plus answers to the questions you actually care about — no formulas,…

Automate a weekly report

Set up a repeatable report once in Claude Desktop, then have Claude regenerate it on demand — so you stop rebuilding it by hand e…

Summarize a folder of documents

Point Claude at a folder of PDFs, notes, or transcripts and get one clear summary — plus answers to questions across all of them.…

Draft documents from a template

Teach Claude your format once in Claude Desktop, then generate consistent, on-brand documents from it again and again.

Desktop actions

Open a folder to work in

Drag it onto Claude, or File → Open

Ask Claude to do something

Type it in the chat, in plain words

Review proposed changes

Read the visual diff in the preview pane

Approve a change

Click Accept on the permission prompt

Undo the last change

Click Undo — originals are kept

Start a new task cleanly

Open a new chat, or clear the current one

Commands

Start Claude Code in your current folder (the terminal / Power-Track way in).

Pick up your most recent conversation right where you left off.

Run one prompt without the chat UI and print the answer — a Power-Track tool for scripts and pipes.

Reopen an earlier conversation from a list of your recent sessions.

Update Claude Code to the latest version.

claude --dangerously-skip-permissions

Start Claude without any permission prompts — a Power-Track flag for fully automated CI and scripts.

Glossary

Software that does things we used to assume needed a human: understanding language, recognizing images, making decisions. It's a…

Machine learning

A way of building software that learns patterns from examples instead of being handed step-by-step rules. Show it thousands of ca…

A Large Language Model — the kind of AI behind Claude. It was trained on a huge amount of text until it became very good at one t…

The moment the model actually runs to produce an answer — its 'thinking' in response to your input. It's the opposite end from tr…

One specific trained AI you can talk to — like Claude Fable 5 or Claude Haiku 4.5. Same idea as car models: one company, differen…

A model that handles more than just text — it can take in images, screenshots, PDFs, diagrams, and sometimes audio, all in the sa…

Course

L00 · The Computer, Explained

Files, folders, apps, paths — the things every later lesson assumes you already know.

L01 · Meet Claude on Desktop

Download the app, say hi in the chat, and learn what Claude actually is — no terminal required.

L02 · The Software World

A short vocabulary tour — client and server, API, database, frontend and backend — so the rest of the course doesn't drown you in…

L03 · Read with Claude

Point Claude at a real codebase and start asking real questions. Files, functions, errors, history — by the end you can open anyt…

L04 · Build with Claude

Your first builds with Claude — point it at your own files, make real changes to real code, read what it proposes, push back when…

L05 · Trust & Safety

Read the prompt, judge the blast, pick the mode, and know what brings you back when something goes wrong.