swift-observation

Each skill follows a consistent format:

```

skill-name/

├── SKILL.md # Entry point with overview and navigation

├── Articles/ # Conceptual guides and tutorials

└── ... # Additional documentation files

```

The SKILL.md file serves as the index with metadata, structure, and links.

---

1. Install uv:

```bash

brew install uv

```

1. Install the [claude-skills-mcp](https://github.com/K-Dense-AI/claude-skills-mcp) server — use the "Add to Cursor" button

1. Create the skills directory:

```bash

mkdir -p ~/.claude/skills

```

1. Clone this repository into the skills folder:

```bash

git clone https://github.com/nonameplum/agent-skills ~/.claude/skills/

```

1. In Cursor chat, type list_skills — first run will download RAG models (may take a moment)

1. Retry after models are downloaded. Skills are now available! 🎉

---

This repository includes swift_docc_to_skill.py — a tool to convert Apple's DocC documentation format into the skill format used here.

Usage

```bash

python swift_docc_to_skill.py [options]

```

Examples

```bash

# Convert Swift language documentation

python swift_docc_to_skill.py TSPL.docc ./programming-swift --skill-name "programming-swift"

# Convert a library's documentation

python swift_docc_to_skill.py Documentation.docc ./my-library --skill-name "my-library"

```

Options

| Option | Description |

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

| --skill-name | Name for the skill (default: derived from input) |

| --skill-description | Description for the skill |

| --main-doc | Specify the main documentation file |

| --skip-wip | Skip work-in-progress files |

| --strip-test-comments | Remove test code comments |

| --skip-assets | Don't copy the Assets directory |

| --dry-run | Preview without making changes |

What It Converts

• → standard markdown links
• @Metadata, @Snippet, @Image → removed/processed
• Double-backtick symbol links → inline code
• Term lists → bold definitions
• Generates SKILL.md with full documentation index