MODULE 01
The AI claim detective
A fact ledger, a revised notice and a four-case evaluation sheet.
YOUR NEIGHBOURHOOD AI WORKSPACE
Learn a little.
Build something useful.
Twelve open modules. Each follows the same loop: learn the foundation, complete a guided build, practise the distinctions and keep your evidence.
All modules are open. Resume remembers your last visit on this browser.
Choose what you want to build.
A suggested sequence gives you a starting point. Explore any module at any time.
Your interactive learning map
Nodes open guided builds. Lines show learning relationships.
Your course: the AI Builder Path
Start with the foundation in each module. The guided build contains the practical steps, copyable prompts and success checks.
01 / Foundations
Learn clear requests and bounded assistant behaviour.
MODULE 02
Weekend Quest: an assistant with boundaries
A role card, a constraint table and two tested itineraries.
02 / Build and remember
Create websites and apps. Link learning notes into a second brain.
MODULE 03
Neighbourhood Swap: your first useful website
A three-event website, a written brief and a browser test checklist.
MODULE 04
BorrowBox: turn a page into a working app
A local BorrowBox app with add, filter, status change and refresh tests.
MODULE 05
Pocket BorrowBox: put your app on a phone
A PWA folder, an install test and an offline capability table.
MODULE 06
Learning Radar: a second brain that explains what to learn next
An Obsidian-compatible learning vault, a cited next-step plan and a missing-evidence test sheet.
03 / Automate and connect
Connect rules, validated AI steps and bot interactions.
MODULE 07
BorrowBox dispatcher: a reliable automation
A workflow export and a six-case input/output table.
MODULE 08
Meetup inbox: AI that produces checkable data
A chain workflow, a schema and a five-case evaluation report.
MODULE 09
A neighbourhood concierge bot
A tested bot, a command table and a webhook troubleshooting note.
04 / Agentic systems
Add bounded tool selection, preferences and human approval.
MODULE 10
The setup planner: an agent you can inspect
An agent workflow, a tool-call trace and a three-case boundary report.
MODULE 11
A writing twin with an honest identity
A consented sample profile, a fact sheet and before/after evaluation pairs.
MODULE 12
Neighbourhood HQ: your personal assistant cockpit
An assistant workflow, test matrix, approved plan and a maintenance runbook.
Resources for your next session
Download the project starter kit · Prepared files and setup · Read the official sources
Open the 36 builder prompts
36 builder prompts: choose a specific task
36 copyable prompts
Draft a 60-word meetup invitation
Write a friendly invitation under 60 words. Fictional practice facts: Saturday, 10:00, community room, free entry. Do not add speakers, food or booking links. Leave missing details blank.Open lesson 01
Simplify an invitation to 40 words
Rewrite my invitation in simpler English, under 40 words. Keep only Saturday, 10:00, community room and free entry. Do not add promises or links. Return one title and the message.Open lesson 01
Check invitation claims against facts
Compare this invitation with the confirmed facts: Saturday, 10:00, community room, free entry. List unsupported claims and missing information. Do not invent replacements. Invitation: [paste draft].Open lesson 01
Plan three tasks after asking for time
Help me plan three fictional tasks: buy stationery, draft a club invitation, tidy the desk. Ask how much time I have before planning. Do not send messages or invent appointments.Open lesson 02
Create a bounded planning assistant
Act as my planning assistant. First ask for tasks, priorities and available time. Use only my answers. Return Task | Next step | Time estimate to confirm. Label assumptions. Draft only; never claim to send, book or complete actions.Open lesson 02
Audit a plan for invented actions
Review this assistant plan against my stated time and priorities. Flag invented tasks, unconfirmed estimates and actions that need my approval. Give one specific correction for each. Plan: [paste].Open lesson 02
Generate a one-file book-club website
Create one complete index.html for a fictional neighbourhood book club. Include a heading, three activities and a sample contact placeholder. Use crimson #E11021, ink #0A0A0B and near-white #FAFAFA. No external assets or forms. Return code only.Open lesson 03
Make the club page responsive
Improve this HTML into a responsive one-page club site. Keep existing facts. Use semantic headings, readable contrast and visible keyboard focus. No tracking, external dependencies or invented links. Explain where to edit the text. HTML: [paste].Open lesson 03
Test the website layout and links
Review my HTML for mobile layout, heading order, contrast, broken links and misleading claims. Give a short test checklist and exact changes. Do not invent test results. HTML: [paste].Open lesson 03
Build an offline task checklist
Create a complete single-file HTML task checklist. Add a task, mark it done and delete it. Use crimson, ink and near-white. Start empty. No network requests, accounts or libraries. Include labels and keyboard controls. Return code only.Open lesson 04
Add saved tasks and JSON export
Add localStorage persistence, Export JSON and Clear data to my checklist. Treat typed text as text, not HTML. Handle empty inputs and damaged stored JSON. No API keys or network calls. Explain how to test on localhost. Code: [paste].Open lesson 04
Review checklist input and storage
Review this checklist code for unsafe HTML insertion, missing labels, empty input handling and persistence failures. Give manual tests, not claimed results. Keep the app offline and single-file. Code: [paste].Open lesson 04
Improve a checklist for phone screens
Make my checklist layout comfortable on a phone: large buttons, clear labels, no horizontal scrolling. Preserve behaviour and local data. No dependencies or invented features. Return the complete HTML. Code: [paste].Open lesson 05
Generate the manifest and service worker
Turn this checklist into a small PWA. Return separate index.html, manifest.json and sw.js. Use supplied icons/192.png and icons/512.png, start_url ./ and standalone display. Version the cache and remove old caches. Cache app files only. Explain localhost and HTTPS tests. Code: [paste].Open lesson 05
Test PWA installation and caching
Review these PWA files for manifest paths, icon sizes, service-worker scope, cache versioning and offline behaviour. Provide browser tests. Do not claim it is installed or app-store ready. Files: [paste].Open lesson 05
Turn a note into linked Markdown
Turn this fictional note into Markdown with Title, Key facts, Open questions and Related notes: Book club meets Saturday at 10:00 in the community room. Entry is free. The organiser has not confirmed refreshments.Open lesson 06
Answer questions with note citations
Answer only from the labelled notes below. Cite note filenames for each factual answer. If the notes do not establish a fact, say Not in my notes. Do not use outside knowledge. Question: [question]. Notes: [paste labelled notes].Open lesson 06
Audit an answer against its notes
Check this answer against the supplied notes. List unsupported claims, incorrect filename citations and missing caveats. Suggest a facts-only revision. Notes: [paste]. Answer: [paste].Open lesson 06
Sketch an urgent-request router
Sketch a tiny workflow: manually start, create a fictional request with urgent=true, send urgent requests to review and others to the normal queue. Name the inputs and the expected output. Do not add AI or external messages.Open lesson 07
Test the router Boolean inputs
Write a manual test table for my request router. It uses a Boolean urgent field: true means review, false means normal. Include a missing-field case and a wrong-type case. Do not claim the workflow was executed.Open lesson 07
Fix missing fields and wrong types
Check my workflow design for missing fields, text-versus-Boolean mistakes and dropped request data. Give one correction at a time. Design: Manual Trigger -> Edit Fields -> If -> review or normal Edit Fields.Open lesson 07
Summarise a request with unknown facts
Summarise this fictional request in one sentence. Use only its facts: I want the community room for a reading circle on Saturday. Time and permission are not confirmed. Mark missing facts as unknown.Open lesson 08
Write the workflow summary prompt
Summarise this request in one sentence, then list missing facts. Use only the input, no promises or invented dates. Request: {{$json.request}}Open lesson 08Check a summary for unsupported facts
Compare the summary with the original request. Identify unsupported facts and missing caveats. Give a corrected summary. Original: [paste]. Summary: [paste].Open lesson 08
Draft an honest test-bot welcome
Design a short factual welcome for a fictional community bot. Say it is a test bot, answers only practice FAQs and is not monitored for urgent requests. Do not invent meetup dates, links or services.Open lesson 09
Write three grounded bot FAQ replies
Write three practice FAQ answers for my test bot using only these fictional facts: free learning group; topics include prompts and automation; the next date is unconfirmed. For other questions say Ask the organiser. No urgent-service claims.Open lesson 09
Review bot promises and missing facts
Review this bot welcome and FAQ for invented information, ambiguous promises and inappropriate urgent-use claims. Keep it clear that this is a workshop test bot. Text: [paste].Open lesson 09
Calculate the sample supplies total
Work out the fictional supplies total: 12 notebooks at 3 units each, plus 8 pens at 2 units each. Show the arithmetic and state the total. This is a workshop calculation, not a real purchase.Open lesson 10
Set rules for a calculator-using agent
You are a workshop supplies assistant. Use Calculator for arithmetic when appropriate. Show the expression and result. Use only supplied quantities. No purchasing, messaging or booking tools are available. Ask for missing quantities; never claim actions were completed.Open lesson 10
Inspect the calculator call and result
Inspect this recorded agent output and tool trace. Did a calculator call occur? Does the result match 12*3+8*2=52? Flag unsupported claims. Do not assume execution from answer text alone. Trace: [paste redacted trace].Open lesson 10
Draft a reminder in an approved style
Draft a fictional club reminder using these style rules: short sentences, warm tone, no hype. Facts: Saturday, 10:00, community room. Label it AI-assisted draft. Do not add promises, names or links.Open lesson 11
Build a labelled style and note assistant
You are a clearly labelled AI style/knowledge assistant, not me. Use only the approved notes and style card below. Cite note names for facts. Say when information is absent. Draft only; never impersonate or claim actions. Style card: [paste]. Notes: [paste].Open lesson 11
Check facts, identity and tone
Compare this draft with my approved style card and notes. Flag invented facts, identity claims and style mismatches. Suggest one revision. Do not infer personality beyond the card. Card: [paste]. Notes: [paste]. Draft: [paste].Open lesson 11
Propose three dependent next actions
Use these fictional notes to suggest three next actions: organise books; ask organiser about room permission; draft an invitation after permission. Mark dependencies. Do not send or book anything. Return Action | Evidence | Approval needed.Open lesson 12
Draft a source-backed weekly plan
Create a weekly plan only from these labelled notes: {{$json.notes}}. Return Action | Evidence note | Dependency | Approval needed. Use at most three actions. Mark missing facts unknown. Produce a draft only; do not claim to send, book or finish work.Open lesson 12Audit plan dependencies and approval
Audit this capstone plan against its notes. Identify invented facts, missing dependencies and actions needing permission. Return Pass/Revise with reasons. Do not approve on my behalf. Notes: [paste]. Plan: [paste].Open lesson 12
Watch the Hyperframes agent-flow example
Watch an agent hand off the work.
Original ten-second Hyperframes render. Follow the request through an agent and a tool, then stop for human review. The lesson diagrams above are crisp vector animations; this is the actual rendered video.
Open Hyperframes MP4Motion pauses when your device requests reduced motion.