Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/majiayu000/spellbooknpx agentmods add skills/majiayu000/spellbook/harmonyos-appWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/majiayu000/spellbook/harmonyos-app)<a href="https://agentmods.dev/skills/majiayu000/spellbook/harmonyos-app"><img src="https://agentmods.dev/badge/skills/majiayu000/spellbook/harmonyos-app/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/majiayu000/spellbook/harmonyos-app"><img src="https://agentmods.dev/badge/skills/majiayu000/spellbook/harmonyos-app.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk pass
- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00043 | $0.02578 |
| Opus 5.5 | $0.00017 | $0.01031 |
| Sonnet 5.5 | $0.00009 | $0.00516 |
| Haiku 4.5 | $0.00004 | $0.00258 |
Grade A, and why
harmonyos-app scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 414 lines — stays where its author put it; the contents beside it link to each section on GitHub.
HarmonyOS Application Development
Core Principles
- ArkTS First — Use ArkTS with strict type safety, no
anyor dynamic types - Declarative UI — Build UI with ArkUI's declarative components and state management
- Stage Model — Use modern Stage model (UIAbility), not legacy FA model
- Distributed by Design — Leverage cross-device capabilities from the start
- Atomic Services — Consider atomic services and cards for lightweight experiences
- One-time Development — Design for multi-device adaptation (phone, tablet, watch, TV)
Hard Rules (Must Follow)
These rules are mandatory. Violating them means the skill is not working correctly.
No Dynamic Types
ArkTS prohibits dynamic typing. Never use any, type assertions, or dynamic property access.
// ❌ FORBIDDEN: Dynamic types
let data: any = fetchData();
let obj: object = {};
obj['dynamicKey'] = value; // Dynamic property access
(someVar as SomeType).method(); // Type assertion
// ✅ REQUIRED: Strict typing
interface UserData {
id: string;
name: string;
}
let data: UserData = fetchData();
// Use Record for dynamic keys
let obj: Record<string, string> = {};
obj['key'] = value; // OK with Record type
No Direct State Mutation
Never mutate @State/@Prop variables directly in nested objects. Use immutable updates.
// ❌ FORBIDDEN: Direct mutation
@State user: User = { name: 'John', age: 25 };
updateAge() {
this.user.age = 26; // UI won't update!
}
// ✅ REQUIRED: Immutable update
updateAge() {
this.user = { ...this.user, age: 26 }; // Creates new object, triggers UI update
}
// For arrays
@State items: string[] = ['a', 'b'];
// ❌ FORBIDDEN
this.items.push('c'); // UI won't update
// ✅ REQUIRED
this.items = [...this.items, 'c'];
Stage Model Only
Always use Stage model (UIAbility). Never use deprecated FA model (PageAbility).
// ❌ FORBIDDEN: FA Model (deprecated)
// config.json with "pages" array
export default {
onCreate() { ... } // PageAbility lifecycle
}
// ✅ REQUIRED: Stage Model
// module.json5 with abilities configuration
import { UIAbility } from '@kit.AbilityKit';
export default class EntryAbility extends UIAbility {
onCreate(want: Want, launchParam: AbilityConstant.LaunchParam): void {
// Modern Stage model lifecycle
}
onWindowStageCreate(windowStage: window.WindowStage): void {
windowStage.loadContent('pages/Index');
}
}
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 414 lines · 43 tokens per session scan A 58bbe64b3071
harmonyos-app is a skill published in the GitHub repository majiayu000/spellbook (288 stars, last pushed today), licensed MIT. It adds 43 tokens to every session and 2,578 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
output-dev-workflow-cost
Calculate and display the cost of an Output SDK workflow execution run. Use when checking LLM token costs, API service costs, or total spend for a specific workflow run.
lov-integrate-lovstudio-llm-skill
An integration with LovStudio.AI’s OpenAI-compatible API for text chat, speech-to-text, speech translation, speech generation, and live audio sessions. It returns structured results that other application code can use.
adhx
Search your own ADHX saves (posts kept from X, Instagram, TikTok, YouTube, Reddit and more) by meaning, including what is in pictures and videos, and fetch any X/Twitter post as clean LLM-friendly JSON. Use when the user asks to find something they saved, bookmarked or watched ("find my saved videos with a frog in…
flowise-builder
Use when flowise visual LLM workflow builder — drag-drop chatflows, API endpoints, document loaders, tools. Use when working with flowise builder.
huggingface-community-evals
Run evaluations for Hugging Face Hub models using inspect-ai and lighteval on local hardware. Use for backend selection, local GPU evals, and choosing between vLLM / Transformers / accelerate. Not for HF Jobs orchestration, model-card PRs, .evalresults publication, or community-evals automation.
sota-llm-engineering
State-of-the-art LLM application engineering rules for BUILDING and AUDITING LLM-powered features. Claude should use this skill whenever it is building, modifying, or reviewing anything that calls a language model — chat features, RAG pipelines, agents and tool use, structured extraction, classification…