agentleFS
DunoDogeDunoDoge/OHOS-Skills/skill-author/SKILL.md

skill-author

Skill1 starsChanged 2 months agoMITView on GitHub

Guide for creating new AI agent skills in this OHOS-Skills repository. Use when the user wants to add a new skill, update an existing skill's SKILL.md, add reference documents, or modify boundary definitions between skills. Triggers include "create a new skill", "add a skill", "update skill description", "add boundary", "write SKILL.md", or any task involving skill authoring conventions in this repo.

23 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 18 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Do not force this skill onto general HarmonyOS development tasks — use the domain skills (arkts-helper, arkts-debug, arkts-build, …

    line 19
  2. Distinctive:

    Before creating a new skill, check whether the topic overlaps with existing skills:

    line 43
  3. Very specific:

    If the new topic fits within an existing skill's scope, extend that skill (add references, update SKILL.md) rather than creating a new one.

    line 58
  4. Very specific:

    After creating a new skill, update the ## Boundary with Other Skills table in all four existing domain skills to include the new skill …

    line 132
  5. Distinctive:

    Add the new skill to the Skills table in README.md.

    line 136

“Common” is measured: kept by at least 6 in 10 of the 18 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

skill-author

Guide for authoring new AI agent skills following the conventions of this repository.

When to Use

  • User wants to create a new skill directory with SKILL.md
  • User wants to update an existing skill's description, workflow, or boundary table
  • User wants to add reference documents to a skill's references/ directory
  • User wants to add code examples to a skill's assets/ directory
  • User asks about skill structure, frontmatter fields, or boundary conventions

Do not force this skill onto general HarmonyOS development tasks — use the domain skills (arkts-helper, arkts-debug, arkts-build, arkts-ndk-dev, harmony-fetch) instead.

Boundary with Other Skills

This skill covers skill authoring conventions, not HarmonyOS development content. When the task involves the following, cross-reference the corresponding skill:

Issue TypeCross-reference SkillBoundary
Writing/reviewing ArkTS codearkts-helperThis skill covers "how to structure a skill"; arkts-helper covers "how to write ArkTS code"
Fixing ArkTS compilation errorsarkts-debugThis skill does not cover error fixes
Diagnosing runtime crashesarkts-crash-diagnosisThis skill does not cover crash diagnosis
Building/deploying HarmonyOS appsarkts-buildThis skill does not cover build commands
NDK/C++ interop developmentarkts-ndk-devThis skill does not cover NDK code
UI/design-language guidanceharmony-designThis skill does not produce design content
App icon design/configcreate-app-iconThis skill does not produce icon assets
Project comprehensionunderstand-hmosThis skill does not explain HarmonyOS projects
Project scaffoldingcreate-hmos-projectThis skill does not create HarmonyOS projects
Fetching online documentationharmony-fetchIf new skill docs are needed from Huawei portal, use harmony-fetch to fetch them first
Validating an existing skillvalidate-skillskill-author creates/edits; validate-skill checks the result

Workflow (Follow Strictly)

1. Determine Skill Scope

Before creating a new skill, check whether the topic overlaps with existing skills:

Existing SkillCovers
arkts-helperArkTS language syntax, style, TS→ArkTS migration, standard library, concurrency, ArkUI state management
arkts-debug28 common ArkTS compilation errors and their fixes
arkts-crash-diagnosisRuntime crash diagnosis: jscrash/faultlog/hilog parsing scripts, fault-mode knowledge base
arkts-buildCommand-line build, test, sign, deploy, CI pipelines
arkts-ndk-devNDK / Node-API / C++ interop development
harmony-designHarmonyOS design language: visual style, layout, motion, components, UX standards
harmony-fetchOnline documentation fetching from Huawei portal (MCP + REST)
create-app-iconApp icon design, layered SVG generation, project config, compliance
understand-hmosProject comprehension: guided tour, file-by-file explanation, config Q&A with verified citations
create-hmos-projectProject scaffolding from a Stage-model template

If the new topic fits within an existing skill's scope, extend that skill (add references, update SKILL.md) rather than creating a new one.

2. Create the Directory Structure

<skill-name>/
  SKILL.md              # Required
  references/           # Required. Must include INDEX.md.
  assets/               # Optional. Code examples (.ets, .cpp, etc.)
  scripts/              # Optional. Utility scripts (e.g., fetch_doc.py)

3. Write SKILL.md

Frontmatter (YAML, required fields):

---
name: <skill-name>
description: <comprehensive trigger condition — list every keyword, error code, API name, and scenario that should activate this skill>
---

The description is critical. The agent uses it to decide when to load the skill. Under-specifying it means the skill won't activate when needed. Include:

  • Specific error codes or messages
  • API names, decorators, classes
  • Scenario keywords
  • Related concepts and synonyms

Body sections (in order):

  1. # Skill Name — one-line purpose statement
  2. ## When to Use — bullet list of activation scenarios
  3. ## Boundary with Other Skills — table with Issue Type | Cross-reference Skill | Boundary. Must reference all four other domain skills where boundaries exist.
  4. ## Workflow (Follow Strictly) — numbered steps the agent must follow

4. Create references/INDEX.md

# INDEX - <skill-name> Reference Documentation

> **Must read first**: Before any <skill-name> task, **read this file first**, then load the corresponding document based on the task type mapping table below.

## Document Roles

| File | Role |
| --- | --- |
| `01-topic.md` | Description of first topic |
| `02-topic.md` | Description of second topic |

## Reading Paths by Scenario

| Scenario | Required Reading |
| --- | --- |
| Scenario A | `01-topic.md` |
| Scenario B | `02-topic.md` → `03-topic.md` |

5. Add Reference Documents

Each doc in references/ must start with:

# Document Title

> **上游 URL**: https://developer.huawei.com/consumer/cn/doc/...
> **抓取时间**: YYYY-MM-DD

[Content follows, using ## / ### heading hierarchy matching upstream]

6. Update Boundary Tables in Other Skills

After creating a new skill, update the ## Boundary with Other Skills table in all four existing domain skills to include the new skill where boundaries exist.

7. Update README.md

Add the new skill to the Skills table in README.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.