in-app-ui
Guidance for implementing Knock in-app UI in a web app, with a focus on setting up, rendering, and debugging Knock guides in React.
What this skill does
# Knock in-app UI skill
This skill helps you build in-app UI with Knock. It covers the two in-app products — **feeds** and **guides** — at a high level, then goes deep on guides: provider setup, rendering with hooks, and debugging.
Reference: https://docs.knock.app/in-app-ui/overview
## Overview
The skill is organized into four focused rule files. Client-framework guidance is scoped per framework via a `-<framework>` suffix (currently only React). Cross-framework concepts live in unsuffixed files.
1. **Feeds vs. guides** (framework-agnostic) — which product to pick for a given surface and why
2. **Setting up the guide providers in React** — `KnockProvider` and `KnockGuideProvider` props, where each value comes from, and how to sequence them
3. **Rendering guides in React** — building a guide component with `useGuide` / `useGuides`, typed content, and engagement tracking
4. **Debugging guides** (framework-agnostic) — the guides toolbar, the triage checklist, and testing workflow
> **Framework scope:** right now this skill only covers React (`@knocklabs/react`). If the user is building with Vue, Svelte, plain JS, React Native, iOS, or Android, stop and ask how they'd like to proceed — do not adapt the React rules to another client SDK on your own.
## How to use this skill
### When deciding what to build
Start with `rules/feeds-vs-guides.md`:
- Confirm the surface you're building is actually a guide, not a feed
- Check the decision table before picking a direction
- If the answer is "both," wrap the app in `KnockProvider` once and render each product's provider where it's needed
### When adding guides to a React app for the first time
1. Read `rules/setup-guide-providers-react.md`
2. **Before running any CLI commands, confirm the CLI is authenticated and which Knock environment this setup is for.** First run `knock whoami` — if it errors with something like "not authenticated" or "no user session," run `knock login` and ask the user to complete the browser flow before continuing (the CLI persists the session so this is a one-time step per machine). Only after `knock whoami` succeeds, run `knock environment list` and ask the user to pick (the CLI defaults to `development`, but most real integrations target `production`). Remember that slug as `<env-slug>` and pass `--environment <env-slug>` on every subsequent environment-scoped `knock` command.
3. **Before asking the user anything about the channel, discover `channelId` via the Knock CLI:** run `knock channel list --json | jq -r '.[] | select(.key == "knock-guide") | .id'`. Channels are account-scoped, so this command does **not** take `--environment`. If it prints a UUID, use it — do not ask the user to confirm or re-paste. Only ask the user if the CLI returns nothing or errors. See the rule file's "Where to get `channelId`" procedure for the full fallback order.
4. Ask the user only for values that can't be auto-discovered — primarily the public `apiKey` for the chosen environment (and confirm `user.id` is coming from the app's auth context). Do not bundle the `apiKey` ask with `channelId`.
5. Wire `KnockProvider` + `KnockGuideProvider` at the top of the tree.
6. Gate `readyToTarget` on any async data your targeting depends on.
7. **Get a real guide rendering before stopping.** Run `knock guide list --environment <env-slug> --json`, show the user the options (`key`, `name`, each step's `schema_key`), and build the first component against a real guide's actual values. Do not scaffold with placeholder strings like `"changelog-card"`. Fetch the message type schema with `knock message-type get <schema_key> --environment <env-slug> --json` so the content is typed. **If the environment has no guides, offer to scaffold a test one via the Knock CLI** (`knock guide new` → edit the JSON → `knock guide push --environment <env-slug>`) using a built-in message type (`card`, `banner`, or `modal`) with obvious-placeholder content — don't stall waiting for manual dashboard setup. See `rules/rendering-guides-react.md` → "First guide: discover real guides via CLI before writing code" for the full procedure including the empty-environment branch.
8. Flag anything that still needs the user (paste `pk_` key, flip the guide to active in the dashboard, restart dev server) explicitly — don't leave them to discover it by absence.
### When building a new guide component in React
1. Follow the workflow in `rules/rendering-guides-react.md`
2. **Discover the target guide (or message type) via the Knock CLI before picking values.** `knock guide list --environment <env-slug> --json` for the guide's `key` / step `schema_key`; `knock message-type get <schema_key> --environment <env-slug> --json` for the content schema. Avoid placeholder strings.
3. Pick `useGuide` for single-guide surfaces, `useGuides` for lists
4. Define a TypeScript type that mirrors the message type schema you just pulled
5. Wire `markAsSeen`, `markAsInteracted`, and `markAsArchived` — custom components must do this themselves
### When a guide isn't rendering
1. Open `rules/debugging-guides.md` and work the triage checklist top to bottom
2. Turn on the guides toolbar (`?knock_guide_toolbar=true`) first — it answers most questions in seconds
3. Distinguish server-side (targeting/eligibility) from client-side (provider/component) failures before digging deeper
## Rule files reference
- `rules/feeds-vs-guides.md` — product selection between feeds and guides (framework-agnostic)
- `rules/setup-guide-providers-react.md` — configuring `KnockProvider` and `KnockGuideProvider` for guides (React)
- `rules/rendering-guides-react.md` — `useGuide`, `useGuides`, typed content, engagement tracking (React)
- `rules/debugging-guides.md` — toolbar, triage checklist, testing workflow (framework-agnostic)
## Quick reference
The examples below are React. For any other client SDK, see the note at the top of **Overview** before proceeding.
### Providers (minimum viable setup — React)
```tsx
<KnockProvider
apiKey={process.env.NEXT_PUBLIC_KNOCK_API_KEY}
user={{ id: currentUser.id }}
>
<KnockGuideProvider
channelId={process.env.NEXT_PUBLIC_KNOCK_GUIDE_CHANNEL_ID}
readyToTarget
listenForUpdates
>
{children}
</KnockGuideProvider>
</KnockProvider>
```
### Where to source each value
- **Auth first, then environment** — Knock is environment-scoped. Before any CLI command, verify the CLI is authenticated with `knock whoami`; if it errors, run `knock login` and wait for the user to complete the browser flow. Then run `knock environment list` and confirm the target slug (`production`, `development`, …) with the user. Pass `--environment <env-slug>` on every subsequent environment-scoped `knock` command. Don't rely on the CLI's `development` default.
- `apiKey` — Knock dashboard → **Platform → API keys** → public `pk_...` key **from the tab for the chosen environment** (switch envs via the dashboard's environment selector first; remind the user to copy the key for the right env)
- `user.id` — your auth context; must match the id used when identifying the user from your backend
- `channelId` — the **UUID** of the guide channel, not its key. Channels are account-scoped, so `knock channel list` does **not** take `--environment`. **Always attempt CLI discovery before asking the user:**
```bash
knock channel list --json | jq -r '.[] | select(.key == "knock-guide") | .id'
```
If this prints a UUID, use it directly — don't prompt for confirmation. The default guide channel key is `knock-guide` (type `in_app_guide`). Fall back to the dashboard (**Settings → Integrations → Channels**) only if the CLI returns nothing or errors. See `rules/setup-guide-providers-react.md` for the full procedure.
### Hooks at a glance
- `useGuide({ type })` — one guide by message type
- `useGuide({ key })` — one specific guide by key
- `useGuides({ type })` — array of guides by message type
- `useGuideContext()` — low-level client access
### Engagement methods
Related in Design
contribute
IncludedLocal-only OSS contribution command center. Auto-refreshes the user's in-flight PR and issue state on invoke so conversations start with full context — no need to brief Claude on what's in flight. Helps the user find issues to contribute to on GitHub, builds per-repo dossiers of what each upstream expects (CLA, DCO, branch convention, AI policy, draft-first, review bots, issue templates), runs deterministic gates before any external action so AI-assisted contributions don't reach maintainers as slop. State is markdown-only: candidate files at ~/.contribute-system/candidates/, repo dossiers at ~/.contribute-system/research/, append-only event log at ~/.contribute-system/log.jsonl. No database, no cloud calls. Use when the user asks about their PRs / issues / contributions, wants to find new work to take on, claim an issue, build/refresh a repo's dossier, or draft a Design Issue or PR. Trigger with "/contribute", "what's my PR status", "find a contribution", "claim issue X", "draft a Design Issue for Y", "refresh dossier for Z".
architectural-analysis
IncludedUser-triggered deep architectural analysis of a codebase or scoped subtree across eight modes — information architecture, data flow, integration points, UI surfaces, interaction patterns, data model, control flow, and failure modes. This skill should be used when the user asks to "diagram this codebase," "map the architecture," "show the data flow," "give me an ERD," "trace control flow," "find the integration points," "verify the layout pattern," "audit the UX architecture," or any similar request whose primary deliverable is mermaid diagrams plus cited reports under docs/architecture/. Dispatches haiku/sonnet sub-agents in parallel for per-mode exploration, then verifies every citation mechanically before any node lands in a diagram. Not for one-off prose explanations of code (use code-explanation) or for high-level system design from scratch (use system-design).
mcp
IncludedModel Context Protocol (MCP) server development and tool management. Languages: Python, TypeScript. Capabilities: build MCP servers, integrate external APIs, discover/execute MCP tools, manage multi-server configs, design agent-centric tools. Actions: create, build, integrate, discover, execute, configure MCP servers/tools. Keywords: MCP, Model Context Protocol, MCP server, MCP tool, stdio transport, SSE transport, tool discovery, resource provider, prompt template, external API integration, Gemini CLI MCP, Claude MCP, agent tools, tool execution, server config. Use when: building MCP servers, integrating external APIs as MCP tools, discovering available MCP tools, executing MCP capabilities, configuring multi-server setups, designing tools for AI agents.
react-native-skia
IncludedDesign, build, debug, and optimise high-polish animated graphics in React Native or Expo using @shopify/react-native-skia, Reanimated, and Gesture Handler. Use when the user wants canvas-driven UI, shaders, paths, rich text, image filters, sprite fields, Skottie, video frames, snapshots, web CanvasKit setup, or performance tuning for custom motion-heavy elements such as loaders, hero art, cards, charts, progress indicators, particle systems, or gesture-driven surfaces. Also use when the user asks for fluid, glow, glass, blob, parallax, 60fps/120fps, or GPU-friendly animated effects in React Native, even if they do not explicitly say "Skia". Do not use for ordinary form/layout work with standard views.
plaid
IncludedProduct Led AI Development — guides founders from idea to launched product. Six capabilities: Idea (discover a product idea), Validate (pressure-test the idea against fatal flaws, problem reality, competition, and 2-week MVP feasibility), Plan (vision intake + document generation), Design (translate image references into a design.md spec), Launch (go-to-market strategy), and Build (roadmap execution). Use when someone says "PLAID", "plaid idea", "help me find an idea", "product idea", "idea from my business", "idea from my expertise", "plaid validate", "validate my idea", "pressure-test", "is this idea good", "find fatal flaws", "validate the problem", "plan a product", "define my vision", "generate a PRD", "product strategy", "plaid design", "design from image", "translate image to design", "create design.md", "extract design tokens", "plaid launch", "go-to-market", "launch plan", "GTM strategy", "launch playbook", "plaid build", "build the app", "start building", or "execute the roadmap".
nextjs-framer-motion-animations
IncludedAdds production-safe Motion for React or Framer Motion animations to Next.js apps, including reveal, hover and tap micro-interactions, whileInView, stagger, AnimatePresence, layout and layoutId transitions, reorder, scroll-linked UI, and lightweight route-content transitions. Use when the user asks to add, refactor, or debug Motion or Framer Motion in App Router or Pages Router codebases, especially around server/client boundaries, reduced motion, LazyMotion, bundle size, hydration, or route transitions. Avoid for GSAP-style timelines, WebGL or 3D scenes, heavy scroll storytelling, or CSS-only effects unless Motion is explicitly requested.