lightning-tv-expert-plugin
Lightning TV framework expertise for @lightningtv/solid Smart TV applications. Auto-triggers on @lightningtv/solid imports. Provides guardra
Open source Open in the app JSON README (API)
About
Lightning TV framework expertise for @lightningtv/solid Smart TV applications. Auto-triggers on @lightningtv/solid imports. Provides guardrails against DOM hallucination and a traversable knowledge graph of 80 atomic API/rule reference notes.
Details
- Kind
- Plugins
- Topic
- AI, RAG & memory
- Publisher
- rwaltenberg
- Origin
- marketplace
- Category
- ferramentas
- Stars
- 3
- Last push
- 2026-04-08T14:56:48Z
- Repository state
- ativo
- License
- MIT
- Added
- 2026-08-30 01:48:58
- Updated
- 2026-08-30 01:48:58
- Origin id
rwaltenberg/lightning-tv-solid-plugin/lightning-tv-expert-plugin
README
# Lightning TV Expert Plugin
A Claude Code plugin that provides framework expertise for building Smart TV applications with [`@lightningtv/solid`](https://github.com/lightning-tv/solid) **v3.x**.
> **Compatibility:** All knowledge was extracted from `@lightningtv/solid` **v3.2.3** source code and is valid for the **v3.x** major version line. APIs may differ in v2.x or earlier.
## How It Works
The plugin installs a **skill** that:
1. **Auto-triggers** when your codebase imports `@lightningtv/solid`
2. **Injects lightweight guardrails** (~300 tokens) into the conversation -- the 6 most critical rules that prevent DOM hallucination
3. **Directs Claude to read detailed references on-demand** from a knowledge graph of 173 atomic entries across 9 domains
Because it's a skill (not an agent), it integrates seamlessly with any workflow -- including subagent-driven development, brainstorming, and TDD. The controller receives the framework constraints and passes them to any subagents it spawns.
## Architecture
```
skills/lightning-tv/
SKILL.md # ~300 tokens: trigger + critical rules + read instructions
notes/
guardrails.md # ~800 tokens: full forbidden tags, DOM APIs, 22-point ruleset
entries/
index.md # Hub: entry point to the knowledge graph
core guide.md # ElementNode, renderer, config, task scheduling
styling guide.md # Flex layout, states, transitions, shaders/effects
focus guide.md # Focus manager, key handling, capture/bubble
navigation guide.md # Row/Column nav, spatial nav, scrolling modes
components guide.md # Row, Column, Grid, Virtual, Image, Lazy, etc.
input guide.md # useFocusManager, useHold, useMouse, key maps
accessibility guide.md # Announcer, speech synthesis, ARIA
performance guide.md # Virtual scrolling, lazy loading, preserve, task queue
routing guide.md # HashRouter, KeepAlive, route preservation
*.md # 173 atomic entries (api, pattern, gotcha, architecture)
```
**Token flow**: trigger costs ~300. First code task reads guardrails (~800). Each domain guide adds ~500-800. Individual entries are ~300-600 each. Claude reads only what's relevant to the current task.
## Knowledge Graph
Every entry is titled as a prose statement and linked via `[[wiki links]]`:
- **77 API entries** -- function signatures, parameters, return values, side effects
- **43 architecture entries** -- how systems work internally, design decisions, data flow
- **34 gotcha entries** -- misleading behaviors, undocumented quirks, silent failures
- **17 pattern entries** -- how to correctly use components and APIs
- **2 comparison entries** -- when to use X vs Y, trade-offs
All entries were extracted directly from `@lightningtv/solid` v3.2.3 source code (55 files) and cross-referenced across 9 domains.
## Installation
```bash
# From the community marketplace
claude plugin marketplace add anthropics/claude-plugins-community
claude plugin install lightning-tv@claude-community
# Or add this project's marketplace directly
claude plugin marketplace add rwaltenberg/lightning-tv-solid-plugin
claude plugin install lightning-tv@lightning-tv-solid-plugin
```
Adding the marketplace directly ensures you always get updates from this repository rather than depending on community marketplace sync.
## Manual Invocation
Type `/lightning-tv` to activate the skill explicitly in any conversation.
## Key Rules Enforced
1. Only `<view>`, `<node>`, `<text>` exist -- no HTML tags
2. No DOM APIs -- no `document`, `addEventListener`, `querySelector`
3. `forwardFocus` required on every container with focusable children
4. Key handlers must `return true` to stop event propagation
5. JSX inline props override `style` object -- use signals or states for dynamics
6. Always set `color={0xffffffff}` when using `src` for images
See `skills/lightning-tv/notes/guardrails.md` for the complete 22-point ruleset.
## Framework Version
| Field | Value |
|-------|-------|
| Framework | `@lightningtv/solid` |
| Version extracted from | **v3.2.3** |
| Compatible major line | **v3.x** |
| Renderer peer dependency | `@lightningjs/renderer ^3.0.0` |
| SolidJS peer dependency | `solid-js *` |
| Router peer dependency | `@solidjs/router ^0.15.1` |
## Compatibility
The skill-based architecture means it works with any Claude Code workflow or plugin. Because the guardrails are injected into the conversation context (not isolated in a subagent), any workflow that dispatches subagents will naturally pass the framework constraints along.
## Source Material
All 173 entries were extracted from the `@lightningtv/solid` v3.2.3 framework source code (55 files), verified against source with cross-referencing across all 9 domains, and checked against a prior v2.0.0 plugin for accuracy (4 incorrect rules corrected, 7 missing gotchas added).