Library / The Full Catalog
Agent-Native OS / The Full Catalog

All 111 use cases.

The bananas list. Real things people in your shoes have built using exactly this stack. The curated 5-stage journey on the Library is one path through this catalog. This page is the rest.

● Live

Once your install is done, you have the operating system. None of these require new tools, new skills, or new subscriptions beyond what's already in your workshop repo. They're all reachable from where you are right now.

The list is intentionally long. Skim it, dog-ear the ones that make you stop and think "wait, I want that," then build one. Items marked LIVE are confirmed running today on the team's daily fleet. The rest are the obvious adjacent unlocks people have built with this same stack in workshops, on YouTube, and in our community.

Browse by section → Download the PDF →
L1 Infrastructure L2 Senses L3 Triggers L4 Autonomy
Patterns

Browse by primary pattern.

Most use cases combine the whole stack: workbench, instructions, tools, secrets, memory, runtime, interfaces, and safety. These groupings show the part of the workflow that carries the most weight so you can find examples by shape.

Briefs and recurring reportsDaily, weekly, and event-triggered summaries that turn many sources into one useful readout. Tool-connected workflowsEmail, calendar, CRM, Slack, Google, APIs, and other services your agent can act through. Memory-backed workflowsTranscripts, notes, source material, and second-brain flows where retrieval quality matters. Human-interface workflowsSlack, Telegram, iMessage, voice, audio, and dashboards where a person receives or steers the work. Scheduled and autonomous workflowsJobs that wake up later, monitor changes, survive delays, or run without a fresh prompt. Review and safety workflowsApprovals, tests, audit trails, checkpointing, and recovery loops. Creative and output workflowsWriting, images, video, audio, scripts, and publishing assets. Business operations workflowsSales, invoices, finance, client work, household ops, and industry-specific workflows.

Pick one that makes you a little nervous.

Use cases below are numbered 1 through 111 continuously across all 13 sections. Pick a number. Open Claude Code in your workshop repo and paste:

"I just finished the workshop. I want to build use case [NUMBER]. Walk me through it."

Then skim the catalog. Dog-ear the ones that make you stop and think "wait, I want that."

01Morning briefings and daily ops

  1. You wake up to a brief on Slack at 8am with calendar, urgent emails, overdue tasks, and the top three "if-you-only-do-one-thing" actions for today.(L4)LIVE
  2. You get the same brief read out loud as a 90-second audio file dropped in your iMessage so you can listen while making coffee.(L4 + voice)
  3. You get a "dad mode" version that strips work, surfaces only family calendar items, school deadlines, and the grocery list for tonight.(L3 + L4)
  4. You get a "kid-friendly" morning brief read aloud during breakfast that covers weather, school events, and one fun fact.(L4 + voice)
  5. You get a sales-focused brief that ranks today's calls by deal size, recent activity, and last-touch date pulled from your CRM.(L2 + L4)
  6. You get an end-of-day brief at 5pm that compares plan-vs-actual, flags promises you made and didn't keep, and queues tomorrow's top three.(L4)
  7. Your brief auto-pulls overnight Granola transcripts and surfaces any decisions or commitments that landed while you were asleep.(L2 + L4)LIVE
  8. You get a Sunday-night "week ahead" brief that reads your calendar and writes a one-paragraph narrative of the upcoming week.(L4)
  9. You get a Monday brief that includes your spouse's calendar so you can coordinate kid pickups before either of you opens a laptop.(L2 + L4)

02Email, calendar, meetings

  1. You auto-process Granola meeting transcripts into structured notes with attendees, decisions, and action items the moment a meeting ends.(L2 + L4)LIVE
  2. You auto-transcribe and file every Zoom recording the moment it lands on disk, including a summary and embeddings for later search.(L4)LIVE
  3. You get draft replies to client emails older than 48 hours, written in your voice, queued in Gmail for one-click send.(L2 + L4)
  4. Calendar holds get auto-scheduled around your "deep work" preferences, and conflicts get a draft decline with a counter-time offered.(L2 + L4)LIVE
  5. You get a Monday morning audit of last week's meetings: which ran long, which had no notes, which produced action items you never closed.(L2 + L4)
  6. Pre-meeting briefs land in your inbox 15 minutes before each call with attendee bios, last-touch summary, and three suggested questions.(L2 + L4)
  7. Post-meeting follow-up emails draft themselves the moment a meeting ends, pulling action items from the transcript.(L2 + L4)
  8. Recurring 1:1s auto-pull a running doc of last meeting's commitments and surface any that slipped.(L4)
  9. You can text "decline tomorrow's 2pm and offer Thursday at 11" and have the agent handle the rescheduling email entirely.(L4 + iMessage)
  10. Your "out of office" detector reads incoming email, identifies vacation auto-replies, and shifts follow-up timing automatically.(L2)

03CRM and lead handling

  1. New leads get auto-enriched with company size, recent funding, LinkedIn highlights, and a draft opening message.(L2 + L4)
  2. Stale opportunities (no touch in 14 days) get a draft "checking in" email tailored to their last conversation.(L2 + L4)
  3. Inbound webform submissions trigger a research agent that posts a one-page brief to Slack before you reply.(L4)
  4. Sales call transcripts auto-extract pain points, budget signals, and competitor mentions into a structured CRM field.(L2 + L4)
  5. You get a daily "follow up with these five people" list ranked by deal value times days-since-last-touch.(L4)
  6. New connections from a conference get auto-categorized, enriched, and queued for personalized 24-hour follow-up.(L2 + L4)
  7. Your pipeline forecast updates itself every Friday by reading recent meeting notes and adjusting close-date confidence.(L4)

04Content and writing

  1. You write five bullet points and get a polished newsletter draft in your voice ready for review.(L3)
  2. Slack posts for cohort sessions get auto-drafted from the prior session's transcript plus the upcoming agenda.(L2 + L3)LIVE
  3. You drop a YouTube link and get a 90-second video script for a follow-up reaction post.(L2 + L3)
  4. You get a weekly content calendar drafted every Monday with three platform-specific post ideas tied to what is trending in your niche.(L2 + L4)
  5. Podcast episode prep packets get assembled the day before each recording with guest bio, recent posts, and five custom questions.(L2 + L3)
  6. Long-form blog posts get a built-in adversarial reviewer agent that flags weak arguments before you publish.(L4)
  7. You give the agent your top five recurring posts and it generates 30 net-new variations every month.(L3)
  8. You get LinkedIn comments drafted (in your voice) for every post in a list of accounts you want to be visible on.(L2 + L4)
  9. Speaking-event proposals draft themselves from a one-paragraph topic plus the conference website.(L2 + L3)
  10. You can paste a half-finished thought into Slack and get back three opening-line variants for a post.(L3)

05Knowledge management

  1. Every meeting you attend gets stored in a searchable graph with attendees, topics, decisions, and links to related meetings.(L2 + L4)LIVE
  2. YouTube videos from 31 monitored channels get auto-transcribed, corrected, and filed nightly with full-text search across years of content.(L4)LIVE
  3. You can ask "what did we decide about X in any meeting in the last six months" and get a cited answer with timestamps.(L2 + L4)
  4. Voice memos auto-transcribe into your second brain with auto-tagging and links to existing notes.(L4)LIVE
  5. Research articles you save get auto-summarized and filed under the project they relate to without you sorting them.(L2 + L4)
  6. Decision logs auto-update when a meeting transcript contains the phrase "let's go with" or "we decided."(L2 + L4)
  7. You can ask the agent "what was that thing someone said about pricing back in February" and get the actual quote.(L4)
  8. Your reading list auto-prioritizes based on what topics you have meetings about this week.(L4)

06Personal life ops

  1. You order the same meal-prep order every Monday by saying "do the brik order," and the agent fills the cart and waits for your confirm.(L4 + iMessage)LIVE
  2. Grocery lists auto-build from the week's meal plan and get sent to Instacart for review.(L4)
  3. Family calendar conflicts trigger a Slack DM between you and your partner with three resolution options.(L2 + L4)
  4. School deadlines (permission slips, picture day, parent-teacher) get extracted from forwarded emails and added to a shared calendar.(L2 + L4)
  5. Household maintenance reminders (HVAC filter, gutter clean, smoke alarm batteries) auto-schedule based on last completion date.(L4)
  6. Birthday and anniversary alerts arrive 7 days out with three gift ideas based on the person's recent interests.(L4)
  7. Vacation prep checklists auto-generate from your destination, length, and weather forecast.(L2 + L4)
  8. Doctor appointment follow-ups (results, refills, next visit) get tracked in one place and surface when overdue.(L2 + L4)
  9. Pet-care reminders (vet visits, medication, food reorder) auto-schedule and reorder via Amazon Subscribe and Save.(L4)

07Creative and media

  1. You generate hyper-realistic infographics for any blog post or social card with one slash command using Nano Banana Pro.(L3)
  2. You create branded session-recap infographics in seconds from a meeting transcript.(L3)
  3. Voice cloning workflows let you produce audio versions of your written content in your own voice for podcast distribution.(L3)LIVE
  4. Video scripts go from rough bullets to teleprompter-ready in one pass with B-roll suggestions.(L3)
  5. You auto-generate carousel posts (6 slides) from a single article with consistent brand styling.(L3)
  6. Image generation pipelines let you produce 30 variations of a hero image with prompted edits, ranked by your previous picks.(L3 + L4)
  7. Director-of-photography pipelines auto-tag and auto-rate every shoot's footage by composition, focus, and subject.(L4)
  8. Music libraries get auto-tagged with mood, BPM, and use-case suggestions for video editors.(L2 + L4)

08Financial and invoices

  1. Invoice tracking auto-detects unpaid invoices over 30 days and drafts polite, escalating follow-up emails.(L2 + L4)
  2. Expense receipts forwarded to a dedicated email get parsed, categorized, and dropped into your bookkeeping system.(L2 + L4)
  3. Stripe payment notifications trigger thank-you emails with onboarding next steps for new customers.(L4)
  4. Monthly P&L summaries get auto-drafted from your accounting export with year-over-year deltas highlighted.(L4)
  5. Cassidy credit usage gets monitored daily with auto-alerts before any student hits their limit, and unblocking is one CLI call.(L4)LIVE
  6. Subscription audits run quarterly and surface every recurring charge over $20/month for your review.(L2 + L4)

09Developer-flavored

  1. PR review bots draft critique comments on every pull request with style, security, and test-coverage notes.(L4)
  2. Deploy notifications post to Slack with a one-line "what changed" summary auto-extracted from the commit messages.(L4)LIVE
  3. Log monitoring agents watch your error stream and ping you only when a new error type appears.(L4)
  4. On-call assistants can run diagnostic commands via Slack DM and walk you through a runbook when paged.(L4 + Slack)
  5. Git checkpoint agents auto-commit working state every N minutes with auto-generated messages so you never lose work.(L4)LIVE
  6. Daily code-review digests summarize what merged across all your repos in the last 24 hours.(L4)
  7. Documentation gaps get auto-detected when new functions land without docstrings, and drafts get queued.(L4)
  8. Schema migrations get a built-in dry-run reviewer that checks for breaking changes before you deploy.(L4)
  9. Test failures trigger a subagent that reproduces the failure locally and proposes a fix as a draft commit.(L4)LIVE
  10. New issues filed on GitHub auto-triage by severity, suggested owner, and proposed labels.(L2 + L4)

10Multi-machine fleet (advanced 8D track)

  1. Heartbeat agents on multiple machines ping each other every 5 minutes and triage what needs human attention.(L4)LIVE
  2. Long-running jobs auto-distribute across your Mac Studio, M1 Max, and Mac Mini based on which has free CPU.(L4)LIVE
  3. Tailscale-meshed agents can hand off work mid-task: one machine starts the research, another finishes the draft.(L4)LIVE
  4. Scheduled jobs survive reboots and crashes via launchd, restarting themselves and logging recovery.(L4)
  5. A central observability dashboard shows every agent run across the fleet with cost, duration, and outcome traces.(L4 + Langfuse)LIVE
  6. Agent-to-agent message queues let one agent file tickets for another to pick up later.(L4)

11Voice, text, iMessage agents

  1. You can text "what's on my plate today" to your agent and get a summary back in 10 seconds, no laptop required.(L4 + iMessage)LIVE
  2. You can text "decline tomorrow's 2pm" and have it handled including the email follow-up.(L4 + iMessage)
  3. Voice-driven journaling: dictate to your agent on a walk and get back a structured journal entry filed by date.(L4 + voice)LIVE
  4. Audio summaries of long Slack threads play to your AirPods on demand.(L4 + voice)
  5. Family group chat assistants help triage who-needs-to-pick-up-which-kid messages without you scrolling.(L4)
  6. You can voice-memo a client follow-up idea and have the agent draft the actual email by the time you sit down.(L4 + voice)
  7. iMessage-based "what should I cook" agents read your fridge inventory (logged via voice) and propose three recipes.(L4 + iMessage)
  8. Group chat summarization: drop into a long unread thread and ask the agent to catch you up in three bullets.(L4)

12Domain-specific power user

  1. Real estate agents auto-build property briefs (comps, school district, walk score, recent sales) from a single MLS link.(L2 + L4)
  2. E-commerce listing managers auto-rewrite product descriptions for SEO and post-cart emails when inventory drops below threshold.(L2 + L4)
  3. Photography pipelines auto-cull shoots, generate proof galleries, and draft client-delivery emails.(L4)
  4. Legal review agents flag standard contract clauses against a personal precedent library and propose redlines.(L2 + L4)
  5. Academic research workflows auto-pull citations, format bibliographies, and surface contradicting papers for any draft.(L2 + L4)
  6. Real-estate investor agents monitor zoning changes, tax lien sales, and FSBO listings in target zip codes.(L2 + L4)
  7. Restaurant operators get nightly summaries of online reviews across Yelp, Google, and TripAdvisor with a draft reply queue.(L2 + L4)
  8. Coaches and consultants get pre-session client briefs that pull last-call notes, recent goals, and "what's changed" signals.(L2 + L4)
  9. Plumbers, electricians, and trades operators get auto-generated bid PDFs from a voice memo describing the job.(L3 + L4)
  10. Etsy and Shopify sellers get auto-drafted seasonal product launches based on Pinterest trend data.(L2 + L4)

13The "wait, I want that" category

  1. You drop a 90-minute meeting recording on the agent and it returns the three most quotable 30-second clips with timestamps.(L4)
  2. The agent reads your last 30 days of email and identifies who you owe responses to but have been ghosting.(L2 + L4)
  3. The agent watches your calendar and warns you on Sunday night if next week is over-scheduled relative to your deep-work targets.(L4)
  4. You can ask "who haven't I talked to in three months that I should reach out to" and get a ranked list with conversation starters.(L2 + L4)
  5. The agent maintains your "second brain" automatically, with no manual filing. Every transcript, voice memo, screenshot, and saved article gets routed to the right note.(L2 + L4)LIVE
  6. You give the agent a half-baked product idea and it returns a competitor scan, market sizing estimate, and three feature MVPs.(L2 + L3)
  7. The agent ghost-writes your weekly recap email to your team in your voice from your actual week's calendar and Slack activity.(L2 + L4)
  8. You can run the same prompt across Claude, Codex, Gemini, and a local Ollama model in parallel and compare answers side-by-side.(L4)
  9. The agent runs a nightly "what changed in my industry" sweep across 20 sources and posts only the genuinely new signal.(L2 + L4)
  10. You build an agent that builds agents. Your /build skill takes a one-sentence problem and returns a deployable workflow.(L4 capstone)

Pick one that makes you a little nervous.

Every one of these has been built with this exact stack. The LIVE ones are running today on the team's daily fleet. The rest are obvious adjacent unlocks people in our community have shipped.

If a use case here makes you stop and say "wait, I want that" — that's your first build.

Open Claude Code in your workshop repo and paste:

"I just finished the workshop. I want to build [USE CASE NUMBER]. Walk me through it."

The agent has the journey, the blueprints, and the templates. You have the operating system. Pick the one that makes you a little nervous to commit to, and build it this week.