mcpbeat

Liveview Patterns

oliver-kriska/claude-elixir-phoenix-opencode-liveview-patterns

Build LiveView: async data (assign_async), PubSub (check connected?), phx-change events, form components/modals/uploads, streams for lists, live_patch. Use when handling interactions, debugging events, or tracking Presence.'

This is a copy. The original lives at oliver-kriska/claude-elixir-phoenix-liveview-patterns.

9k tokens
context cost
the whole folder, loaded on every use
7
files
instructions only
0
copies elsewhere
how many repositories repackaged it
514
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/oliver-kriska/claude-elixir-phoenix --skill liveview-patterns

What comes with it

29 962 bytes besides the instruction
references/async-streams.md
references/channels-presence.md
references/components.md
references/forms-uploads.md
references/js-interop.md
references/pubsub-navigation.md

The instruction itself

12 sections, as written by the author

LiveView Patterns Reference

> Ash projects: Use ash-framework skill for AshPhoenix.Form. Lifecycle: AshPhoenix.Form.validate/3 on phx-change, AshPhoenix.Form.submit/2 on submit, to_form/1 for HEEx. Do not use Ecto.Changeset.cast/3.

Reference for building with Phoenix LiveView 1.0/1.1.

Iron Laws — Never Violate These

  • NO UNCONDITIONAL DB QUERIES IN MOUNT — Mount runs TWICE. Default: assign_async. SEO routes: connected? guard + cache-backed disconnected branch (crawlers read that HTML)
  • ALWAYS USE STREAMS FOR LISTS — Regular assigns = O(n) memory per user. Streams = O(1)
  • CHECK connected?/1 BEFORE SUBSCRIPTIONS — Prevents double subscriptions
  • EXTRACT VARIABLES BEFORE assign_async CLOSURE — Closures copy entire referenced variables
  • LOAD PRIMARY DATA IN mount/3, PAGINATION IN handle_params/3 — handle_params runs on EVERY URL change
  • NEVER PASS SOCKET TO BUSINESS LOGIC — Extract data before calling contexts
  • CHECK CHANGESET ERRORS BEFORE UI DEBUGGING — Silent form save = check {:error, changeset} first, not viewport/JS
  • HIDDEN INPUTS FOR ALL REQUIRED EMBEDDED FIELDS — Every required field in an embedded schema MUST have a hidden_input if not directly editable
  • NEVER USE assign_new FOR LIFECYCLE VALUESassign_new skips the function if key exists. Use assign/3 for locale, current user, or any value refreshed every mount

10. MATCH {:error, %Ecto.Changeset{}} EXPLICITLY — Bare {:error, _} merges changeset and non-changeset errors; the form silently never re-renders validation errors. Handle other errors separately

Memory Impact

| Pattern | 3K items | 10K users × 10K items |

|---------|----------|----------------------|

| Regular assigns | ~5.1 MB | ~10+ GB |

| Streams | ~1.1 MB | Minimal (O(1)) |

Decision: Lists with >100 items → Use streams, not assigns

Quick Patterns

Async Assigns (CRITICAL)

def mount(%{"slug" => slug}, _session, socket) do
  # Extract needed values BEFORE the closure
  scope = socket.assigns.current_scope

  {:ok,
   socket
   |> assign_async(:org, fn -> {:ok, %{org: fetch_org(scope, slug)}} end)}
end

Streams for Lists

def mount(_params, _session, socket) do
  {:ok, stream(socket, :items, Items.list_items())}
end

# Insert/update/delete
stream_insert(socket, :items, item, at: 0)
stream_delete(socket, :items, item)

SEO Dead-Render (cache-backed disconnected branch)

For public/SEO-visible routes (marketing, articles, product listings) the

disconnected render IS the HTML crawlers see. Fetch from a cache there, real

data on connect:

def mount(_params, _session, socket) do
  products =
    if connected?(socket),
      do: Catalog.list_products(),
      else: Cache.get_products() || []

  {:ok, assign(socket, products: products)}
end

Empty list → <noscript>-friendly skeleton. Cache → :persistent_term, ETS,

or Cachex. This satisfies Iron Law #1 AND keeps Googlebot/GPTBot happy.

PubSub with connected? check

def mount(_params, _session, socket) do
  if connected?(socket), do: Chat.subscribe(room_id)
  {:ok, socket}
end
Same LiveView, different params? → patch / push_patch
Different LiveView, same live_session? → navigate / push_navigate
Different live_session or non-LiveView? → href / redirect

Component Decision Tree

Does component need BOTH internal state AND event handling?
│
├── YES → Does it encapsulate APPLICATION logic (not just DOM)?
│   ├── YES → Use LiveComponent ✅
│   └── NO → Refactor to function component with parent handling
│
└── NO → Use Function Component ✅

Official guidance: "Prefer function components over live components"

Common Anti-patterns

| Wrong | Right |

|-------|-------|

| DB queries without assign_async | Use assign_async for all queries |

| assign(socket, items: list) for lists | stream(socket, :items, list) |

| PubSub subscribe without connected? | if connected?(socket), do: subscribe() |

| Passing socket to context functions | Extract socket.assigns first |

| Business logic in handle_event | Delegate to context |

| assign_new for locale/user in hooks | assign/3 (must run every mount) |

References

For detailed patterns, see:

  • references/async-streams.md - assign_async, stream_async, streams
  • references/forms-uploads.md - Forms, validation, file uploads
  • references/components.md - Function components, LiveComponents
  • references/pubsub-navigation.md - PubSub, navigation, JS commands
  • references/js-interop.md - Third-party JS libraries, phx-update="ignore", hooks
  • references/channels-presence.md - Phoenix Channels, Presence, token auth

How to use it

Copy the folder

Take oliver-kriska/claude-elixir-phoenix-opencode-liveview-patterns from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.